Research Progress and Prospect of Cultivated Land Extraction from Remote Sensing Images
Obtaining real-time and accurate information on cultivated land distribution is a critical task in modern land resource management and high-quality agricultural development.With the rapid development of satellite technology,remote sensing monitoring has gradually become an important method for extracting culti-vated land information.At the same time,deep learning technology has risen rapidly and is increasingly be-coming a pivotal technique for cultivated land extraction from remote sensing imagers.In this paper,recent re-search findings on cultivated land extraction both domestically and internationally were summarized,the limita-tions of traditional extraction algorithms,and the positive significance of high-resolution remote sensing images to cultivated land extraction were elaborated.The fundamental processes of cultivated land extraction were out-lined,and the primary development stages and research strategies of cultivated land extraction algorithms were reviewed,and the main optimization methods for such algorithms,and the application of multi-task network models were summarized.Finally,the limitations of current deep learning algorithms were discussed and future trends in the development of cultivated land extraction technologies were anticipated.
Cultivated land extractionDeep learningSemantic segmentationHigh resolution image
国家科技期刊平台
NSTL
万方数据
