A simple, efficient, and practical approach for detecting cloud and shadow areas in satellite imagery and restoring them with clean pixel values has been developed. Cloud and shadow areas are detected using spectral information from the blue, shortwave infrared, and thermal infrared bands of Landsat Thematic Mapper or Enhanced Thematic Mapper Plus imagery from two dates (a target image and a reference image). These detected cloud and shadow areas are further refined using an integration process and a false shadow removal process according to the geometric relationship between cloud and shadow. Cloud and shadow filling is based on the concept of the Spectral Similarity Group (SSG), which uses the reference image to find similar alternative pixels in the target image to serve as replacement values for restored areas. Pixels are considered to belong to one SSG if the pixel values from Landsat bands 3, 4, and 5 in the reference image are within the same spectral ranges. This new approach was applied to five Landsat path/rows across different landscapes and seasons with various types of cloud patterns. Results show that almost all of the clouds were captured with minimal commission errors, and shadows were detected reasonably well. Among five test scenes, the lowest producer's accuracy of cloud detection was 93.9% and the lowest user's accuracy was 89%. The overall cloud and shadow detection accuracy ranged from 83.6% to 99.3%. The pixel-filling approach resulted in a new cloud-free image that appears seamless and spatially continuous despite differences in phenology between the target and reference images. Our methods offer a straightforward and robust approach for preparing images for the new 2011 National Land Cover Database production.
We developed a method that analyzes the quality of the cultivated cropland class mapped in the USA National Land Cover Database (NLCD) 2006. The method integrates multiple geospatial datasets and a Multi Index Integrated Change Analysis (MIICA) change detection method that captures spectral changes to identify the spatial distribution and magnitude of potential commission and omission errors for the cultivated cropland class in NLCD 2006. The majority of the commission and omission errors in NLCD 2006 are in areas where cultivated cropland is not the most dominant land cover type. The errors are primarily attributed to the less accurate training dataset derived from the National Agricultural Statistics Service Cropland Data Layer dataset. In contrast, error rates are low in areas where cultivated cropland is the dominant land cover. Agreement between model-identified commission errors and independently interpreted reference data was high (79%). Agreement was low (40%) for omission error comparison. The majority of the commission errors in the NLCD 2006 cultivated crops were confused with low-intensity developed classes, while the majority of omission errors were from herbaceous and shrub classes. Some errors were caused by inaccurate land cover change from misclassification in NLCD 2001 and the subsequent land cover post-classification process.
An accurate and historical land cover monitoring dataset for Alaska could provide fundamental information for a range of studies, such as conservation habitats, biogeochemical cycles, and climate systems, in this distinctive region. This research addresses challenges associated with the extraction of training data for timely and accurate land cover classifications in Alaska over longer time periods (e.g., greater than 10 years). Specifically, we designed the “Region-Specific Model Adaptation (RSMA)” method for training data. The method integrates land cover information from the National Land Cover Database (NLCD), LANDFIRE’s Existing Vegetation Type (EVT), and the National Wetlands Inventory (NWI) and machine learning techniques to generate robust training samples based on the Anderson Level II classification legend. The assumption of the method is that spectral signatures vary across regions because of diverse land surface compositions; however, despite these variations, there are consistent, collective land cover characteristics that span the entire region. Building upon this assumption, this research utilized the classification power of deep learning algorithms and the generalization ability of RSMA to construct a model for the RSMA method. Additionally, we interpreted existing vegetation plot information for land cover labels as validation data to reduce inconsistency in the human interpretation. Our validation results indicate that the RSMA method improved the quality of the training data derived solely from the NLCD by approximately 30% for the overall accuracy. The validation assessment also demonstrates that the RSMA method can generate reliable training data on large scales in regions that lack sufficient reliable data.