Land surface temperature (LST) retrieved from satellite thermal sensors often consists of mixed t... more Land surface temperature (LST) retrieved from satellite thermal sensors often consists of mixed temperature components. Retrieving subpixel LST is therefore needed in various environmental and ecological studies. In this paper, we developed two methods for downscaling coarse resolution thermal infrared (TIR) radiance for the purpose of subpixel temperature retrieval. The first method was developed on the basis of a scale-invariant physical model on TIR radiance. The second method was based on a statistical relationship between TIR radiance and land cover fraction at high spatial resolution. The two methods were applied to downscale simulated 990-m ASTER TIR data to 90-m resolution. When validated against the original 90-m ASTER TIR data, the results revealed that both downscaling methods were successful in capturing the general patterns of the original data and resolving considerable spatial details. Further quantitative assessments indicated a strong agreement between the true values and the estimated values by both methods.
Using a linear unconstrained least squares (LSS) method and a non-linear artificial neural networ... more Using a linear unconstrained least squares (LSS) method and a non-linear artificial neural network (ANN) algorithm, we conducted a spectral mixture analysis to the Advanced Spaceborne Thermal Emission and Reflectance Radiometer (ASTER) image data in Yokohama city, Japan, for mapping the abundance of the urban surface components. ASTER is a newly developed research facility instrument. The regions of interest of four endmembers (Vegetation, Soil, High/Low albedo impervious surfaces) were determined in Maximum Noise Fraction (MNF) feature spaces. The spectral signatures of the four endmembers were then extracted from the ASTER VNIR (15-m resolution) and SWIR (30-m resolution) imagery by referring to high spatial resolution airborne imagery (The Airborne Imaging Spectrometer, AISA, with 2-m resolution) and land use/land cover map for training and testing the LSS and ANN algorithms. Experimental results indicate that ASTER VNIR and SWIR image data are capable of mapping the abundances of urban surface components with a reasonable accuracy and that the ANN outperforms the unconstrained LSS in this spectral mixture analysis.
Data from three thermal sensors with different spatial resolution were assessed for urban surface... more Data from three thermal sensors with different spatial resolution were assessed for urban surface temperature retrieval over the Yokohama City, Japan. The sensors are Thermal Airborne Broadband Imager (TABI), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and MODerate resolution Imaging Spectroradiometer (MODIS). Two algorithms were developed for land surface temperature (LST) retrieval from TABI image and ASTER thermal infrared (TIR) channels 13 and 14. In addition, ASTER LST and MODIS LST products were also collected. All the LST images were assessed by analyzing the relationship between LST and normalized difference vegetation index (NDVI) and by spatial distributions of LST profiles, derived from typical transects over the LST images. In this study, a strong negative relationship between LST and NDVI has been demonstrated although the degree of correlation between NDVI and LST varies slightly among the different LST images. Cross-validation among the LST images retrieved from the three thermal sensors of different spatial resolutions indicates that the LST images retrieved from the 2 channel ASTER data and a single band TABI thermal image using our developed algorithms are reliable. The LST images retrieved from the three sensors should have different potential to urban environmental studies. The MODIS thermal sensor can be used for the synoptic overview of an urban area and for studying urban thermal environment. The ASTER, with its TIR subsystem of 90-m resolution, allows for a more accurate determination of thermal patterns and properties of urban land use/land cover types, and hence, a more accurate determination of the LST. In consideration of the high heterogeneity of urban environment, the TABI thermal image, with a high spatial resolution of 2 m, can be used for rendering and assessing complex urban thermal patterns and detailed distribution of LST at the individual house level more accurately.
Land surface temperature (LST) is a key parameter in numerous environmental studies. Surface hete... more Land surface temperature (LST) is a key parameter in numerous environmental studies. Surface heterogeneity induces uncertainty in estimating subpixel temperature. To take an advantage of simultaneous, multiresolution observations at coincident nadirs by the Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER) and the MODerate-resolution Imaging Spectroradiometer (MODIS), LST products from the two sensors were examined for a portion of suburb area in Beijing, China. We selected Soil-Adjusted Vegetation Index (SAVI), Normalized Multi-band Drought Index (NMDI), Normalized Difference Built-up Index (NDBI) and Normalized Difference Water Index (NDWI) as representative remote sensing indices for four land cover types (vegetation, bare soil, impervious and water area), respectively. By using support vector machines, the overall classification accuracy of the four land cover types with inputs of the four remote sensing indices, extracted from ASTER visible near infrared (VNIR) bands and shortwave infrared (SWIR) bands, reached 97.66%, and Kappa coefficient was 0.9632. In order to lower the subpixel temperature estimation error caused by re-sampling of remote sensing data, a disaggregation method for subpixel temperature using the remote sensing endmember index based technique (DisEMI) was established in this study. Firstly, the area ratios and statistical information of endmember remote sensing indices were calculated from ASTER VNIR/SWIR data at 990 m and 90 m resolutions, respectively. Secondly, the relationship between the 990 m resolution MODIS LST and the corresponding input parameters (area ratios and endmember indices at the 990 m resolution) was trained by a genetic algorithm and self-organizing feature map artificial neural network (GA-SOFM-ANN). Finally, the trained models were employed to estimate the 90 m resolution subpixel temperature with inputs of area ratios and endmember indices at the 90 m resolution. ASTER LST product was used for verifying the estimated subpixel temperature, and the verified results indicate that the estimated temperature distribution was basically consistent with that of ASTER LST product. A better agreement was found between temperatures derived by our proposed method (DisEMI) and the ASTER 90 m data (R 2 = 0.709 and RMSE = 2.702 K).
Fires in boreal and temperate forests play a significant role in the global carbon cycle. While f... more Fires in boreal and temperate forests play a significant role in the global carbon cycle. While forest fires in North America (NA) have been surveyed extensively by U.S. and Canadian forest services, most fire records are limited to seasonal statistics without information on temporal evolution and spatial expansion. Such dynamic information is crucial for modeling fire emissions. Using the daily Advanced Very High Resolution Radiometer (AVHRR) data archived from 1989 to 2000, an extensive and consistent fire product was developed across the entire NA forest regions on a daily basis at 1-km resolution. The product was generated following data calibration, geo-referencing, and the application of an active fire detection algorithm and a burned area mapping algorithm. The spatial-temporal variation of forest fire in NA is analyzed in terms of (1) annual and monthly patterns of fire occurrences in different eco-domains, (2) the influence of topographic factors (elevation zones, aspect classes, and slope classes), and (3) major forest types and eco-regions in NA. It was found that 1) among the 12 years analyzed, 1989 and 1995 were the most severe fire years in NA; 2) the majority of burning occurred during June-July and in low elevation zones (b 500 m) with gentle slopes (b 10°), except in the dry eco-domain where more fires occurred in higher elevation zones (N 2000 m); 3) most fires occurred in the polar ecodomain, sub-arctic eco-division, and in the taiga ( boreal forests), forest-tundras and open woodlands eco-provinces in the boreal forests of Canada. The tendency for multiple burns to occur increases with elevation and slope until about 2500 m elevation and 24°slope, and decreases therefore. In comparison with ground observations, the omission and commission errors are on the order of 20%.
The invasive weed yellow starthistle (Centaurea solstitialis) has infested between 4 and 6 millio... more The invasive weed yellow starthistle (Centaurea solstitialis) has infested between 4 and 6 million hectares in California. It often forms dense infestations and rapidly depletes soil moisture, preventing the establishment of other species. Precise assessment of its canopy cover, especially low-density abundance in the earlier growing season, is the key to effective management. Compact Airborne Spectrographic Imager 2 (CASI-2) hyperspectral imagery was acquired at the western edge of California's Central Valley grasslands on July 15, 2003. Four linear spectral mixture models (LSMM) were investigated from the original CASI-2 data. Band selections based upon residual analysis and feature extraction (PCA) were further explored to reduce the data dimension. All approaches, except four band-selection unconstrained LSMMs, provide consistent results. The uncertainty of the PCA-based LSMM was estimated through a Monte-Carlo simulation. The maximum standard deviation was approximately 11%. The results suggest that unmixing CASI-2 imagery could be used for estimating and mapping yellow starthistle for larger regional areas.
A comparison of the performance of three feature extraction methods was made for mapping forest c... more A comparison of the performance of three feature extraction methods was made for mapping forest crown closure (CC) and leaf area index (LAI) with EO-1 Hyperion data. The methods are band selection (SB), principal component analysis (PCA) and wavelet transform (WT). Hyperion data were acquired on October 9, 2001. A total of 38 field measurements of CC and LAI were collected on August 10 -11, 2001, at Blodgett Forest Research Station, University of California at Berkeley, USA. The analysis method consists of (1) conducting atmospheric correction with High Accuracy Atmospheric Correction for Hyperspectral Data (HATCH) to retrieve surface reflectance, (2) extracting features with the three methods: SB, PCA and WT, (3) establishing multivariate regression prediction models, (4) predicting and mapping pixel-based CC and LAI values, and (5) validating the CC and LAI mapped results with photo-interpreted CC and LAI values. The experimental results indicate that the energy features extracted by the WT method are the most effective for mapping forest CC and LAI (mapped accuracy (MA) for CC = 84.90%, LAI MA = 75.39%), followed by the PCA method (CC MA = 77.42%, LAI MA = 52.36%). The SB method performed the worst (CC MA = 57.77%, LAI MA = 50.87%).
Abstract Remote sensing (RS) techniques have been widely considered to be a prom-ising source of ... more Abstract Remote sensing (RS) techniques have been widely considered to be a prom-ising source of information for land management decisions. The objective of this study was to develop and compare different methods of delineating management zones (MZs) in a field of winter ...
Powdery mildew is one of the most serious diseases that have a significant impact on the producti... more Powdery mildew is one of the most serious diseases that have a significant impact on the production of winter wheat. As an effective alternative to traditional sampling methods, remote sensing can be a useful tool in disease detection. This study attempted to use multi-temporal moderate resolution satellite-based data of surface reflectances in blue (B), green (G), red (R) and near infrared (NIR) bands from HJ-CCD (CCD sensor on Huanjing satellite) to monitor disease at a regional scale. In a suburban area in Beijing, China, an extensive field campaign for disease intensity survey was conducted at key growth stages of winter wheat in 2010. Meanwhile, corresponding time series of HJ-CCD images were acquired over the study area. In this study, a number of single-stage and multi-stage spectral features, which were sensitive to powdery mildew, were selected by using an independent t-test. With the selected spectral features, four advanced methods: mahalanobis distance, maximum likelihood classifier, partial least square regression and mixture tuned matched filtering were tested and evaluated for their performances in disease mapping. The experimental results showed that all four algorithms could generate disease maps with a generally correct distribution pattern of powdery mildew at the grain filling stage (Zadoks 72). However, by comparing these disease maps with ground survey data (validation samples), all of the four algorithms also produced a variable degree of error in estimating the disease occurrence and severity. Further, we found that the integration of MTMF and PLSR algorithms could result in a significant accuracy improvement of identifying and determining the disease intensity (overall accuracy of 72% increased to 78% and kappa coefficient of 0.49 increased to 0.59). The experimental results also demonstrated that the multi-temporal satellite images have a great potential in crop diseases mapping at a regional scale.
A Compact Airborne Spectrographic Imager-2 (CASI) dataset was used for detecting mortality and ve... more A Compact Airborne Spectrographic Imager-2 (CASI) dataset was used for detecting mortality and vegetation stress associated with a new forest disease. We first developed a multilevel classification scheme to improve classification accuracy. Then, the CASI raw data were transformed to reflectance and corrected for topography, and a principal component (PC) transformation of all 48 bands and the visible bands and NIR bands were separately conducted to extract features from the CASI data. Finally, we classified the calibrated and corrected CASI imagery using a maximum likelihood classifier and tested the relative accuracies of classification across the scheme. The multilevel scheme consists of four levels (Levels 0 to 3). Level 0 covered the entire study area, classifying eight classes (oak trees, California bay trees, shrub areas, grasses, dead trees, dry areas, wet areas, and water). At Level 1, the vegetated and non-vegetated areas were separated. The vegetated and nonvegetated areas were further subdivided into four vegetated (oak trees, California bay trees, shrub areas, grasses) and four non-vegetated (dead trees, dry areas, wet areas, and water) classes at Level 2. Level 3 identified stressed and non-stressed oak trees (two classes). The ten classes classified at different levels are defined as final classes in this study. The experimental results indicated that classification accuracy generally increased as the detailed classification level increased. When the CASI topographically corrected reflectance data were processed into ten PCs (five PCs from the visible region and five PCs from NIR bands), the classification accuracy for Level 2 vegetated classes (non-vegetated classes) increased to 80.15 percent (94.10 percent) from 78.07 percent (92.66 percent) at Level 0. The accuracy of separating stressed from non-stressed oak trees at Level 3 was 75.55 percent. When classified as a part of Level 0, the stressed and non-stressed were almost inseparable. Furthermore, we found that PCs derived from visible and NIR bands separately yielded more accurate results than the PCs from all 48 CASI bands.
Change-vector analysis (CVA) is a valuable technique for landuse/land-cover change detection. How... more Change-vector analysis (CVA) is a valuable technique for landuse/land-cover change detection. However, how to reasonably determine thresholds of change magnitude and change direction is a bottleneck to its proper application. In this paper, a new method is proposed to improve CVA. The method (the improved CVA) consists of two stages, Double-Window Flexible Pace Search (DFPS), which aims at determining the threshold of change magnitude, and direction cosines of change vectors for determining change direction (category) that combines single-date image classification with a minimum-distance categorizing technique. When the improved CVA was applied to the detection of the land-use/land-cover changes in the Haidian District, Beijing, China, Kappa coefficients of "change/ no-change" detection and "from-to" types of change detection were 0.87 and greater than 0.7, respectively, for all kinds of land-use changes. The experimental results indicate that the improved CVA has good potential in land-use/land-cover change detection.
ABSTRACT Public sector policies regulating land development can have a major influence on the env... more ABSTRACT Public sector policies regulating land development can have a major influence on the environmental impacts of urban development, yet few empirical studies have examined the impact of these policies. Our study attempted to address this gap by examining the relationship between a land development code associated with the protection of trees and the extent of urban tree cover on residential parcels. We developed an accurate very high resolution urban tree cover classification using IKONOS imagery, quantified parcel-specific tree cover and evaluated the relationship between year of construction, adoption of a tree ordinance and extent of canopy cover in Tampa, Florida, USA and in nearby areas lacking similar regulation. Statistical results revealed significantly greater tree cover on parcels with homes built after compared to before adoption of tree protection policies, despite a trend toward increased building cover. After controlling for the effects of spatial dependence and relevant physical and socio-demographic characteristics at the census block, multivariate regression results indicate the percentage of homes built after implementing the policy was a strong predictor of increased tree cover. These results and comparisons with nearby areas further suggest Tampa's 1974 tree protection regulations appears to have been effective at increasing tree cover, but other influential factors were also at work. This study demonstrates the potential for evaluating the impact of specific land development policies using remote sensing and other spatial data analysis techniques. In addition, we argue that scientific evidence of the effectiveness of past policy should be used a guide for the creation of future land development regulation.
1] This paper presents an evaluation of advanced very high resolution radiometer (AVHRR)-based re... more 1] This paper presents an evaluation of advanced very high resolution radiometer (AVHRR)-based remote sensing algorithms for detecting active vegetation fires and mapping burned areas throughout North America. The procedures were originally designed for application in Canada with AVHRR data aboard the NOAA 14 satellite. They were tested here with both NOAA 11 and NOAA 14 covering the period 1989-2000. It was found that the active fire detection algorithm performs well with low commission and omission error rates over forested regions in the absence of cloud cover. Moderate errors were found over semi-arid areas covered by thin clouds, as well as along rivers and around lakes observed from sun-glint angles. A modification to a fire algorithm threshold and the addition of a new test can significantly improve the detection accuracy. Burned areas mapped by satellite were compared against extensive fire polygon data acquired by U.S. forest agencies in five western states. The satellite-based mapping matches nearly 90% of total forested burned area, with the difference being mainly attributable to omission of some nonburned islands and patches within the fire polygons. In addition, it maps a significant area of burning outside the fire polygons that appear to be true fires. The 10% omission error was found to be caused mainly by three factors: lack or insufficient number of active fires, partial burning, and vegetation recovery after early season burning. In addition to total area, the location and shapes of burned scars are consistent with the ground-based maps. Overall, the two algorithms are competent for detecting and mapping forest fires in North America north of Mexico with minor modifications.
Improvement in remote sensing techniques in spatial/spectral resolution strengthens their applica... more Improvement in remote sensing techniques in spatial/spectral resolution strengthens their applicability for urban environmental study. Unfortunately, high spatial resolution imagery also increases internal variability in land cover units and can cause a 'salt-and-pepper' effect, resulting in decreased accuracy using pixel-based classification results. Region-based classification techniques, using an image object (IO) rather than a pixel as a classification unit, appear to hold promise as a method for overcoming this problem. Using IKONOS high spatial resolution imagery, we examined whether the IO technique could significantly improve classification accuracy compared to the pixel-based method when applied to urban land cover mapping in Tampa Bay, FL, USA. We further compared the performance of an artificial neural network (ANN) and a minimum distance classifier (MDC) in urban detailed land cover classification and evaluated whether the classification accuracy was affected by the number of extracted IO features. Our analysis methods included IKONOS image data calibration, data fusion with the pansharpening (PS) process, Hue-Intensity-Saturation (HIS) transferred indices and textural feature extraction, and feature selection using a stepwise discriminant analysis (SDA). The classification results were evaluated with visually interpreted data from high-resolution (0.3 m) digital aerial photographs. Our results indicate a statistically significant difference in classification accuracy between pixel-and object-based techniques; ANN outperforms MDC as an objectbased classifier; and the use of more features (27 vs. 9 features) increases the IO classification accuracy, although the increase is statistically significant for the MDC but not for the ANN.
In this study, mixed coniferous forest crown closure (CC) and leaf area index (LAI) were measured... more In this study, mixed coniferous forest crown closure (CC) and leaf area index (LAI) were measured at the Blodgett Forest Research Station, University of California at Berkeley, USA. Data from EO-1 Hyperspectral Imager (Hyperion) and Advanced Land Imager (ALI) acquired on 9 October 2001, and from Landsat 7 Enhanced Thematic Mapper Plus (ETM+) on 25 October 2001 were used for estimation of CC and LAI. A total of 38 forest CC and LAI measurements were used in this correlation analysis. The analysis procedure consists of (1) atmospheric correction to retrieve surface reflectance from Hyperion, ALI and ETM+ data, (2) a total of 38 patches, corresponding to ground CC and LAI measurement plots, extracted from data from the three sensors, and (3) calculating univariate/multivariate correlation coefficient (R 2 ) and root mean square error (RMSE) using CC and LAI measurements and retrieved surface reflectance data of the three sensors. The experimental results indicate: (1) higher individual band correlations with CC and LAI appear in visible and short wave infrared (SWIR) regions due to spectral absorption features (pigments in visible and water and other biochemicals in SWIR); (2) based on ALI individual band wavelengths, the R 2 /RMSE produced with Hyperion bands are all better than those with ALI, except ALI band 1, due to atmospheric scattering of Hyperion bands in the visible region; (3) based on ETM+ individual band wavelengths, Hyperion is better than ALI, which is better than ETM+, especially for the NIR band group of Hyperion; (4) based on spectral region, Hyperion, again, is better than ALI which is better than ETM+, and optimal results appear in the visible region for ALI and in SWIR for Hyperion; and (5) if considering just six bands or six features (six principal components) in estimating CC and LAI, optimal results are obtained with six bands selected from the 167 Hyperion bands. In general, for estimation of forest CC and LAI in this study, the Hyperion sensor has outperformed the ALI and ETM+ sensors, whereas ALI is better than ETM+. The best spectral region for Hyperion is SWIR, but for ALI and ETM+, the visible region should be considered instead.
A total of 139 reflectance spectra (between 350 and 2500 nm) from coast live oak (Quercus agrifol... more A total of 139 reflectance spectra (between 350 and 2500 nm) from coast live oak (Quercus agrifolia) leaves were measured in the laboratory with a spectrometer FieldSpecAPro FR. Correlation analysis was conducted between absorption features, three-band ratio indices derived from the spectra and corresponding relative water content (RWC, %) of oak leaves. The experimental results indicate that there exist linear relationships between the RWC of oak leaves and absorption feature parameters: wavelength position (WAVE), absorption feature depth (DEP), width (WID) and the multiplication of DEP and WID (AREA) at the 975 nm, 1200 nm and 1750 nm positions and two three-band ratio indices: RATIO 975
We developed a method for integrated analysis of multi-source data for vegetation classification ... more We developed a method for integrated analysis of multi-source data for vegetation classification at the continental scale, and applied it to China. Multi-temporal 1 km NOAA Advanced Very High Resolution Radiometer (AVHRR) Holdridge's life zone system and its vegetation-climate classification indices such as bio-temperature (BT), potential evapotranspiration rate (PER) and precipitation (P) correspond better with undisturbed vegetation types all over the world. We generated 1 km images of BT, PER and P using the quantitative model of Holdridge's life zone system with climate data of China. They were processed with principal component analysis (PCA) to produce an ancillary image. This image and 12 monthly images of maximum Normalized Difference Vegetation Index (NDVI) values at 1 km resolution were input into an ISODATA clustering algorithm to carry out a vegetation classification. As a result, 47 information classes were obtained. Seasonal NDVI parameters derived through time series analysis (TSA) of the NDVI temporal profile and a set of quantitative vegetation-climate parameters of Holdrige's life zone model were synthetically utilized to label information classes. In this method, climate, terrain and spectral data were integrated; separability between vegetation types and classification accuracy were improved. A total of 47 land cover classes were obtained. Validation data collected in the field using GPS indicated that an overall classification accuracy of 71.4% was reached, an 8.1% improvement to the map derived only from multi-temporal NDVI images. To compare our results with the International Geosphere-Biosphere Programme (IGBP) DISCover land cover dataset, we aggregated our land cover classes according to the IGBP classification system. The overall classification accuracy for the aggregated vegetation map from our classification results improved IGBP land cover map from 75.5% to 86.3%.
... You are signed in as: Google (Institutional account). Group of crawlers (Institutional accoun... more ... You are signed in as: Google (Institutional account). Group of crawlers (Institutional account ...
Using an unconstrained least squares solution (LSS) method and an artificial neural network (ANN)... more Using an unconstrained least squares solution (LSS) method and an artificial neural network (ANN) algorithm, we estimated oakwood crown closure from a Landsat Thematic Mapper (TM) image of Tulare County, California, USA. Fractions of endmembers (oak crown (f1), grass (f2) and soil (f3)) from mixed pixels were derived from aerial photographs (scale 1 : 40 000) scanned at 1 m ground resolution for training and testing the LSS and ANN algorithms. The aerial photographs were orthorectified using a digital photogrammetric software package with ground control points collected through a differential global positioning system (GPS). The TM image was georeferenced with respect to the corresponding orthorectified aerial photographs. The training and test samples were randomly selected from the TM image and their corresponding fractions of endmembers were derived from the orthophoto. A fourth endmember, shade (f4), was directly extracted from the TM image. Experimental results indicate that the ANN has performed better than the unconstrained LSS. To extract oakwood crown closure in mixed pixels, better results were obtained without using a shade endmember.
Land surface temperature (LST) retrieved from satellite thermal sensors often consists of mixed t... more Land surface temperature (LST) retrieved from satellite thermal sensors often consists of mixed temperature components. Retrieving subpixel LST is therefore needed in various environmental and ecological studies. In this paper, we developed two methods for downscaling coarse resolution thermal infrared (TIR) radiance for the purpose of subpixel temperature retrieval. The first method was developed on the basis of a scale-invariant physical model on TIR radiance. The second method was based on a statistical relationship between TIR radiance and land cover fraction at high spatial resolution. The two methods were applied to downscale simulated 990-m ASTER TIR data to 90-m resolution. When validated against the original 90-m ASTER TIR data, the results revealed that both downscaling methods were successful in capturing the general patterns of the original data and resolving considerable spatial details. Further quantitative assessments indicated a strong agreement between the true values and the estimated values by both methods.
Using a linear unconstrained least squares (LSS) method and a non-linear artificial neural networ... more Using a linear unconstrained least squares (LSS) method and a non-linear artificial neural network (ANN) algorithm, we conducted a spectral mixture analysis to the Advanced Spaceborne Thermal Emission and Reflectance Radiometer (ASTER) image data in Yokohama city, Japan, for mapping the abundance of the urban surface components. ASTER is a newly developed research facility instrument. The regions of interest of four endmembers (Vegetation, Soil, High/Low albedo impervious surfaces) were determined in Maximum Noise Fraction (MNF) feature spaces. The spectral signatures of the four endmembers were then extracted from the ASTER VNIR (15-m resolution) and SWIR (30-m resolution) imagery by referring to high spatial resolution airborne imagery (The Airborne Imaging Spectrometer, AISA, with 2-m resolution) and land use/land cover map for training and testing the LSS and ANN algorithms. Experimental results indicate that ASTER VNIR and SWIR image data are capable of mapping the abundances of urban surface components with a reasonable accuracy and that the ANN outperforms the unconstrained LSS in this spectral mixture analysis.
Data from three thermal sensors with different spatial resolution were assessed for urban surface... more Data from three thermal sensors with different spatial resolution were assessed for urban surface temperature retrieval over the Yokohama City, Japan. The sensors are Thermal Airborne Broadband Imager (TABI), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and MODerate resolution Imaging Spectroradiometer (MODIS). Two algorithms were developed for land surface temperature (LST) retrieval from TABI image and ASTER thermal infrared (TIR) channels 13 and 14. In addition, ASTER LST and MODIS LST products were also collected. All the LST images were assessed by analyzing the relationship between LST and normalized difference vegetation index (NDVI) and by spatial distributions of LST profiles, derived from typical transects over the LST images. In this study, a strong negative relationship between LST and NDVI has been demonstrated although the degree of correlation between NDVI and LST varies slightly among the different LST images. Cross-validation among the LST images retrieved from the three thermal sensors of different spatial resolutions indicates that the LST images retrieved from the 2 channel ASTER data and a single band TABI thermal image using our developed algorithms are reliable. The LST images retrieved from the three sensors should have different potential to urban environmental studies. The MODIS thermal sensor can be used for the synoptic overview of an urban area and for studying urban thermal environment. The ASTER, with its TIR subsystem of 90-m resolution, allows for a more accurate determination of thermal patterns and properties of urban land use/land cover types, and hence, a more accurate determination of the LST. In consideration of the high heterogeneity of urban environment, the TABI thermal image, with a high spatial resolution of 2 m, can be used for rendering and assessing complex urban thermal patterns and detailed distribution of LST at the individual house level more accurately.
Land surface temperature (LST) is a key parameter in numerous environmental studies. Surface hete... more Land surface temperature (LST) is a key parameter in numerous environmental studies. Surface heterogeneity induces uncertainty in estimating subpixel temperature. To take an advantage of simultaneous, multiresolution observations at coincident nadirs by the Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER) and the MODerate-resolution Imaging Spectroradiometer (MODIS), LST products from the two sensors were examined for a portion of suburb area in Beijing, China. We selected Soil-Adjusted Vegetation Index (SAVI), Normalized Multi-band Drought Index (NMDI), Normalized Difference Built-up Index (NDBI) and Normalized Difference Water Index (NDWI) as representative remote sensing indices for four land cover types (vegetation, bare soil, impervious and water area), respectively. By using support vector machines, the overall classification accuracy of the four land cover types with inputs of the four remote sensing indices, extracted from ASTER visible near infrared (VNIR) bands and shortwave infrared (SWIR) bands, reached 97.66%, and Kappa coefficient was 0.9632. In order to lower the subpixel temperature estimation error caused by re-sampling of remote sensing data, a disaggregation method for subpixel temperature using the remote sensing endmember index based technique (DisEMI) was established in this study. Firstly, the area ratios and statistical information of endmember remote sensing indices were calculated from ASTER VNIR/SWIR data at 990 m and 90 m resolutions, respectively. Secondly, the relationship between the 990 m resolution MODIS LST and the corresponding input parameters (area ratios and endmember indices at the 990 m resolution) was trained by a genetic algorithm and self-organizing feature map artificial neural network (GA-SOFM-ANN). Finally, the trained models were employed to estimate the 90 m resolution subpixel temperature with inputs of area ratios and endmember indices at the 90 m resolution. ASTER LST product was used for verifying the estimated subpixel temperature, and the verified results indicate that the estimated temperature distribution was basically consistent with that of ASTER LST product. A better agreement was found between temperatures derived by our proposed method (DisEMI) and the ASTER 90 m data (R 2 = 0.709 and RMSE = 2.702 K).
Fires in boreal and temperate forests play a significant role in the global carbon cycle. While f... more Fires in boreal and temperate forests play a significant role in the global carbon cycle. While forest fires in North America (NA) have been surveyed extensively by U.S. and Canadian forest services, most fire records are limited to seasonal statistics without information on temporal evolution and spatial expansion. Such dynamic information is crucial for modeling fire emissions. Using the daily Advanced Very High Resolution Radiometer (AVHRR) data archived from 1989 to 2000, an extensive and consistent fire product was developed across the entire NA forest regions on a daily basis at 1-km resolution. The product was generated following data calibration, geo-referencing, and the application of an active fire detection algorithm and a burned area mapping algorithm. The spatial-temporal variation of forest fire in NA is analyzed in terms of (1) annual and monthly patterns of fire occurrences in different eco-domains, (2) the influence of topographic factors (elevation zones, aspect classes, and slope classes), and (3) major forest types and eco-regions in NA. It was found that 1) among the 12 years analyzed, 1989 and 1995 were the most severe fire years in NA; 2) the majority of burning occurred during June-July and in low elevation zones (b 500 m) with gentle slopes (b 10°), except in the dry eco-domain where more fires occurred in higher elevation zones (N 2000 m); 3) most fires occurred in the polar ecodomain, sub-arctic eco-division, and in the taiga ( boreal forests), forest-tundras and open woodlands eco-provinces in the boreal forests of Canada. The tendency for multiple burns to occur increases with elevation and slope until about 2500 m elevation and 24°slope, and decreases therefore. In comparison with ground observations, the omission and commission errors are on the order of 20%.
The invasive weed yellow starthistle (Centaurea solstitialis) has infested between 4 and 6 millio... more The invasive weed yellow starthistle (Centaurea solstitialis) has infested between 4 and 6 million hectares in California. It often forms dense infestations and rapidly depletes soil moisture, preventing the establishment of other species. Precise assessment of its canopy cover, especially low-density abundance in the earlier growing season, is the key to effective management. Compact Airborne Spectrographic Imager 2 (CASI-2) hyperspectral imagery was acquired at the western edge of California's Central Valley grasslands on July 15, 2003. Four linear spectral mixture models (LSMM) were investigated from the original CASI-2 data. Band selections based upon residual analysis and feature extraction (PCA) were further explored to reduce the data dimension. All approaches, except four band-selection unconstrained LSMMs, provide consistent results. The uncertainty of the PCA-based LSMM was estimated through a Monte-Carlo simulation. The maximum standard deviation was approximately 11%. The results suggest that unmixing CASI-2 imagery could be used for estimating and mapping yellow starthistle for larger regional areas.
A comparison of the performance of three feature extraction methods was made for mapping forest c... more A comparison of the performance of three feature extraction methods was made for mapping forest crown closure (CC) and leaf area index (LAI) with EO-1 Hyperion data. The methods are band selection (SB), principal component analysis (PCA) and wavelet transform (WT). Hyperion data were acquired on October 9, 2001. A total of 38 field measurements of CC and LAI were collected on August 10 -11, 2001, at Blodgett Forest Research Station, University of California at Berkeley, USA. The analysis method consists of (1) conducting atmospheric correction with High Accuracy Atmospheric Correction for Hyperspectral Data (HATCH) to retrieve surface reflectance, (2) extracting features with the three methods: SB, PCA and WT, (3) establishing multivariate regression prediction models, (4) predicting and mapping pixel-based CC and LAI values, and (5) validating the CC and LAI mapped results with photo-interpreted CC and LAI values. The experimental results indicate that the energy features extracted by the WT method are the most effective for mapping forest CC and LAI (mapped accuracy (MA) for CC = 84.90%, LAI MA = 75.39%), followed by the PCA method (CC MA = 77.42%, LAI MA = 52.36%). The SB method performed the worst (CC MA = 57.77%, LAI MA = 50.87%).
Abstract Remote sensing (RS) techniques have been widely considered to be a prom-ising source of ... more Abstract Remote sensing (RS) techniques have been widely considered to be a prom-ising source of information for land management decisions. The objective of this study was to develop and compare different methods of delineating management zones (MZs) in a field of winter ...
Powdery mildew is one of the most serious diseases that have a significant impact on the producti... more Powdery mildew is one of the most serious diseases that have a significant impact on the production of winter wheat. As an effective alternative to traditional sampling methods, remote sensing can be a useful tool in disease detection. This study attempted to use multi-temporal moderate resolution satellite-based data of surface reflectances in blue (B), green (G), red (R) and near infrared (NIR) bands from HJ-CCD (CCD sensor on Huanjing satellite) to monitor disease at a regional scale. In a suburban area in Beijing, China, an extensive field campaign for disease intensity survey was conducted at key growth stages of winter wheat in 2010. Meanwhile, corresponding time series of HJ-CCD images were acquired over the study area. In this study, a number of single-stage and multi-stage spectral features, which were sensitive to powdery mildew, were selected by using an independent t-test. With the selected spectral features, four advanced methods: mahalanobis distance, maximum likelihood classifier, partial least square regression and mixture tuned matched filtering were tested and evaluated for their performances in disease mapping. The experimental results showed that all four algorithms could generate disease maps with a generally correct distribution pattern of powdery mildew at the grain filling stage (Zadoks 72). However, by comparing these disease maps with ground survey data (validation samples), all of the four algorithms also produced a variable degree of error in estimating the disease occurrence and severity. Further, we found that the integration of MTMF and PLSR algorithms could result in a significant accuracy improvement of identifying and determining the disease intensity (overall accuracy of 72% increased to 78% and kappa coefficient of 0.49 increased to 0.59). The experimental results also demonstrated that the multi-temporal satellite images have a great potential in crop diseases mapping at a regional scale.
A Compact Airborne Spectrographic Imager-2 (CASI) dataset was used for detecting mortality and ve... more A Compact Airborne Spectrographic Imager-2 (CASI) dataset was used for detecting mortality and vegetation stress associated with a new forest disease. We first developed a multilevel classification scheme to improve classification accuracy. Then, the CASI raw data were transformed to reflectance and corrected for topography, and a principal component (PC) transformation of all 48 bands and the visible bands and NIR bands were separately conducted to extract features from the CASI data. Finally, we classified the calibrated and corrected CASI imagery using a maximum likelihood classifier and tested the relative accuracies of classification across the scheme. The multilevel scheme consists of four levels (Levels 0 to 3). Level 0 covered the entire study area, classifying eight classes (oak trees, California bay trees, shrub areas, grasses, dead trees, dry areas, wet areas, and water). At Level 1, the vegetated and non-vegetated areas were separated. The vegetated and nonvegetated areas were further subdivided into four vegetated (oak trees, California bay trees, shrub areas, grasses) and four non-vegetated (dead trees, dry areas, wet areas, and water) classes at Level 2. Level 3 identified stressed and non-stressed oak trees (two classes). The ten classes classified at different levels are defined as final classes in this study. The experimental results indicated that classification accuracy generally increased as the detailed classification level increased. When the CASI topographically corrected reflectance data were processed into ten PCs (five PCs from the visible region and five PCs from NIR bands), the classification accuracy for Level 2 vegetated classes (non-vegetated classes) increased to 80.15 percent (94.10 percent) from 78.07 percent (92.66 percent) at Level 0. The accuracy of separating stressed from non-stressed oak trees at Level 3 was 75.55 percent. When classified as a part of Level 0, the stressed and non-stressed were almost inseparable. Furthermore, we found that PCs derived from visible and NIR bands separately yielded more accurate results than the PCs from all 48 CASI bands.
Change-vector analysis (CVA) is a valuable technique for landuse/land-cover change detection. How... more Change-vector analysis (CVA) is a valuable technique for landuse/land-cover change detection. However, how to reasonably determine thresholds of change magnitude and change direction is a bottleneck to its proper application. In this paper, a new method is proposed to improve CVA. The method (the improved CVA) consists of two stages, Double-Window Flexible Pace Search (DFPS), which aims at determining the threshold of change magnitude, and direction cosines of change vectors for determining change direction (category) that combines single-date image classification with a minimum-distance categorizing technique. When the improved CVA was applied to the detection of the land-use/land-cover changes in the Haidian District, Beijing, China, Kappa coefficients of "change/ no-change" detection and "from-to" types of change detection were 0.87 and greater than 0.7, respectively, for all kinds of land-use changes. The experimental results indicate that the improved CVA has good potential in land-use/land-cover change detection.
ABSTRACT Public sector policies regulating land development can have a major influence on the env... more ABSTRACT Public sector policies regulating land development can have a major influence on the environmental impacts of urban development, yet few empirical studies have examined the impact of these policies. Our study attempted to address this gap by examining the relationship between a land development code associated with the protection of trees and the extent of urban tree cover on residential parcels. We developed an accurate very high resolution urban tree cover classification using IKONOS imagery, quantified parcel-specific tree cover and evaluated the relationship between year of construction, adoption of a tree ordinance and extent of canopy cover in Tampa, Florida, USA and in nearby areas lacking similar regulation. Statistical results revealed significantly greater tree cover on parcels with homes built after compared to before adoption of tree protection policies, despite a trend toward increased building cover. After controlling for the effects of spatial dependence and relevant physical and socio-demographic characteristics at the census block, multivariate regression results indicate the percentage of homes built after implementing the policy was a strong predictor of increased tree cover. These results and comparisons with nearby areas further suggest Tampa's 1974 tree protection regulations appears to have been effective at increasing tree cover, but other influential factors were also at work. This study demonstrates the potential for evaluating the impact of specific land development policies using remote sensing and other spatial data analysis techniques. In addition, we argue that scientific evidence of the effectiveness of past policy should be used a guide for the creation of future land development regulation.
1] This paper presents an evaluation of advanced very high resolution radiometer (AVHRR)-based re... more 1] This paper presents an evaluation of advanced very high resolution radiometer (AVHRR)-based remote sensing algorithms for detecting active vegetation fires and mapping burned areas throughout North America. The procedures were originally designed for application in Canada with AVHRR data aboard the NOAA 14 satellite. They were tested here with both NOAA 11 and NOAA 14 covering the period 1989-2000. It was found that the active fire detection algorithm performs well with low commission and omission error rates over forested regions in the absence of cloud cover. Moderate errors were found over semi-arid areas covered by thin clouds, as well as along rivers and around lakes observed from sun-glint angles. A modification to a fire algorithm threshold and the addition of a new test can significantly improve the detection accuracy. Burned areas mapped by satellite were compared against extensive fire polygon data acquired by U.S. forest agencies in five western states. The satellite-based mapping matches nearly 90% of total forested burned area, with the difference being mainly attributable to omission of some nonburned islands and patches within the fire polygons. In addition, it maps a significant area of burning outside the fire polygons that appear to be true fires. The 10% omission error was found to be caused mainly by three factors: lack or insufficient number of active fires, partial burning, and vegetation recovery after early season burning. In addition to total area, the location and shapes of burned scars are consistent with the ground-based maps. Overall, the two algorithms are competent for detecting and mapping forest fires in North America north of Mexico with minor modifications.
Improvement in remote sensing techniques in spatial/spectral resolution strengthens their applica... more Improvement in remote sensing techniques in spatial/spectral resolution strengthens their applicability for urban environmental study. Unfortunately, high spatial resolution imagery also increases internal variability in land cover units and can cause a 'salt-and-pepper' effect, resulting in decreased accuracy using pixel-based classification results. Region-based classification techniques, using an image object (IO) rather than a pixel as a classification unit, appear to hold promise as a method for overcoming this problem. Using IKONOS high spatial resolution imagery, we examined whether the IO technique could significantly improve classification accuracy compared to the pixel-based method when applied to urban land cover mapping in Tampa Bay, FL, USA. We further compared the performance of an artificial neural network (ANN) and a minimum distance classifier (MDC) in urban detailed land cover classification and evaluated whether the classification accuracy was affected by the number of extracted IO features. Our analysis methods included IKONOS image data calibration, data fusion with the pansharpening (PS) process, Hue-Intensity-Saturation (HIS) transferred indices and textural feature extraction, and feature selection using a stepwise discriminant analysis (SDA). The classification results were evaluated with visually interpreted data from high-resolution (0.3 m) digital aerial photographs. Our results indicate a statistically significant difference in classification accuracy between pixel-and object-based techniques; ANN outperforms MDC as an objectbased classifier; and the use of more features (27 vs. 9 features) increases the IO classification accuracy, although the increase is statistically significant for the MDC but not for the ANN.
In this study, mixed coniferous forest crown closure (CC) and leaf area index (LAI) were measured... more In this study, mixed coniferous forest crown closure (CC) and leaf area index (LAI) were measured at the Blodgett Forest Research Station, University of California at Berkeley, USA. Data from EO-1 Hyperspectral Imager (Hyperion) and Advanced Land Imager (ALI) acquired on 9 October 2001, and from Landsat 7 Enhanced Thematic Mapper Plus (ETM+) on 25 October 2001 were used for estimation of CC and LAI. A total of 38 forest CC and LAI measurements were used in this correlation analysis. The analysis procedure consists of (1) atmospheric correction to retrieve surface reflectance from Hyperion, ALI and ETM+ data, (2) a total of 38 patches, corresponding to ground CC and LAI measurement plots, extracted from data from the three sensors, and (3) calculating univariate/multivariate correlation coefficient (R 2 ) and root mean square error (RMSE) using CC and LAI measurements and retrieved surface reflectance data of the three sensors. The experimental results indicate: (1) higher individual band correlations with CC and LAI appear in visible and short wave infrared (SWIR) regions due to spectral absorption features (pigments in visible and water and other biochemicals in SWIR); (2) based on ALI individual band wavelengths, the R 2 /RMSE produced with Hyperion bands are all better than those with ALI, except ALI band 1, due to atmospheric scattering of Hyperion bands in the visible region; (3) based on ETM+ individual band wavelengths, Hyperion is better than ALI, which is better than ETM+, especially for the NIR band group of Hyperion; (4) based on spectral region, Hyperion, again, is better than ALI which is better than ETM+, and optimal results appear in the visible region for ALI and in SWIR for Hyperion; and (5) if considering just six bands or six features (six principal components) in estimating CC and LAI, optimal results are obtained with six bands selected from the 167 Hyperion bands. In general, for estimation of forest CC and LAI in this study, the Hyperion sensor has outperformed the ALI and ETM+ sensors, whereas ALI is better than ETM+. The best spectral region for Hyperion is SWIR, but for ALI and ETM+, the visible region should be considered instead.
A total of 139 reflectance spectra (between 350 and 2500 nm) from coast live oak (Quercus agrifol... more A total of 139 reflectance spectra (between 350 and 2500 nm) from coast live oak (Quercus agrifolia) leaves were measured in the laboratory with a spectrometer FieldSpecAPro FR. Correlation analysis was conducted between absorption features, three-band ratio indices derived from the spectra and corresponding relative water content (RWC, %) of oak leaves. The experimental results indicate that there exist linear relationships between the RWC of oak leaves and absorption feature parameters: wavelength position (WAVE), absorption feature depth (DEP), width (WID) and the multiplication of DEP and WID (AREA) at the 975 nm, 1200 nm and 1750 nm positions and two three-band ratio indices: RATIO 975
We developed a method for integrated analysis of multi-source data for vegetation classification ... more We developed a method for integrated analysis of multi-source data for vegetation classification at the continental scale, and applied it to China. Multi-temporal 1 km NOAA Advanced Very High Resolution Radiometer (AVHRR) Holdridge's life zone system and its vegetation-climate classification indices such as bio-temperature (BT), potential evapotranspiration rate (PER) and precipitation (P) correspond better with undisturbed vegetation types all over the world. We generated 1 km images of BT, PER and P using the quantitative model of Holdridge's life zone system with climate data of China. They were processed with principal component analysis (PCA) to produce an ancillary image. This image and 12 monthly images of maximum Normalized Difference Vegetation Index (NDVI) values at 1 km resolution were input into an ISODATA clustering algorithm to carry out a vegetation classification. As a result, 47 information classes were obtained. Seasonal NDVI parameters derived through time series analysis (TSA) of the NDVI temporal profile and a set of quantitative vegetation-climate parameters of Holdrige's life zone model were synthetically utilized to label information classes. In this method, climate, terrain and spectral data were integrated; separability between vegetation types and classification accuracy were improved. A total of 47 land cover classes were obtained. Validation data collected in the field using GPS indicated that an overall classification accuracy of 71.4% was reached, an 8.1% improvement to the map derived only from multi-temporal NDVI images. To compare our results with the International Geosphere-Biosphere Programme (IGBP) DISCover land cover dataset, we aggregated our land cover classes according to the IGBP classification system. The overall classification accuracy for the aggregated vegetation map from our classification results improved IGBP land cover map from 75.5% to 86.3%.
... You are signed in as: Google (Institutional account). Group of crawlers (Institutional accoun... more ... You are signed in as: Google (Institutional account). Group of crawlers (Institutional account ...
Using an unconstrained least squares solution (LSS) method and an artificial neural network (ANN)... more Using an unconstrained least squares solution (LSS) method and an artificial neural network (ANN) algorithm, we estimated oakwood crown closure from a Landsat Thematic Mapper (TM) image of Tulare County, California, USA. Fractions of endmembers (oak crown (f1), grass (f2) and soil (f3)) from mixed pixels were derived from aerial photographs (scale 1 : 40 000) scanned at 1 m ground resolution for training and testing the LSS and ANN algorithms. The aerial photographs were orthorectified using a digital photogrammetric software package with ground control points collected through a differential global positioning system (GPS). The TM image was georeferenced with respect to the corresponding orthorectified aerial photographs. The training and test samples were randomly selected from the TM image and their corresponding fractions of endmembers were derived from the orthophoto. A fourth endmember, shade (f4), was directly extracted from the TM image. Experimental results indicate that the ANN has performed better than the unconstrained LSS. To extract oakwood crown closure in mixed pixels, better results were obtained without using a shade endmember.
Uploads
Papers by Ruiliang Pu