研究目的
Investigating the effectiveness of a novel frequency-based classification framework and algorithm (Wave-CLASS) using an overcomplete decomposition procedure for identifying detailed urban land cover types in high spatial resolution data.
研究成果
The Wave-CLASS algorithm is effective in identifying detailed urban land cover types in high spatial resolution data, achieving higher overall accuracies than the maximum-likelihood classifier. It demonstrates robustness for accurate LULC classification with similar producers' and users' accuracies for all classes.
研究不足
The study is limited to high spatial resolution data (QuickBird) and may require adaptation for other data types. The effectiveness of the algorithm on smaller or more spectrally similar objects needs further investigation.
1：Experimental Design and Method Selection:
The study employs an overcomplete wavelet decomposition approach for image classification, comparing it with the maximum-likelihood classifier.
2：Sample Selection and Data Sources:
Three image subsets of QuickBird data over a central region in the city of Phoenix were used.
3：List of Experimental Equipment and Materials:
QuickBird satellite data with four channels (blue, green, red, and near-infrared).
4：Experimental Procedures and Operational Workflow:
Training samples were selected for seven land cover classes, and the Wave-CLASS algorithm was applied to classify the images.
5：Data Analysis Methods:
The effectiveness of the Wave-CLASS algorithm was evaluated using overall accuracy, producer's accuracy, user's accuracy, and kappa coefficient.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
