陳蘇晉，孟令奎，張  文

(武漢大學遙感信息工程學院，湖北 武漢 430079)

　　摘  要：為服務于河湖“清四亂”行動，加強河長對“亂占”問題的管理，以圍湖造田和坑塘養殖2種“亂占”現象為研究對象，提出基于綜合指數法、灰度共生矩陣及支持向量機等多種方法的“亂占”目標識別機制。運用高分辨率遙感影像，分別使用指數法和灰度共生矩陣提取“亂占”目標的光譜和紋理特征，并在融合光譜與紋理特征的基礎上，利用支持向量機進行監督分類，輸出混淆矩陣分析分類精度。研究結果表明，該方法應用于圍湖造田和坑塘養殖2類“亂占”目標識別效果良好，能夠有效減少實地調研工作量，為各級河長監督整改河湖“四亂”問題提供技術支撐。

　　關鍵詞：遙感監測；河長制；河湖“四亂”；特征提取；目標識別

Research on telemetry monitoring technology for “unauthorized rivers and lakes occupation” phenomenon under river chief system

CHEN Sujin, MENG Lingkui, ZHANG Wen

(School of Remote Sensing and Information Engineering, Wuhan University,Wuhan 430079, China)

　　Abstract: To serve “sweeping four unauthorizations” of rivers and lakes action and strengthen river chiefs’ management against on the problem of unauthorized occupation, this paper proposes the unauthorized occupation objects identification mechanism based on several methods such as comprehensive index, Gray Level Co-occurrence Matrix(GLCM) and Support Vector Machine(SVM) method, studying on two unauthorized occupation phenomena: land reclamation and pond aquaculture. Using high resolution remote sensing images, the spectral and texture features of unauthorizedly occupied targets are extracted by using index method and GLCM. By fusing spectral and texture features, SVM is used for supervising classification. Confusion matrix is output to analyze classification accuracy. Result shows that the method does well in unauthorizedly occupied targets identification of land reclamation and pond aquaculture. The method can effectively reduce the workload of field research and provide river chiefs technical support for supervision and remediation of the four unauthorizations problems.

　　Key words: telemetry monitoring； river chief system；rivers and lakes “four unauthorizations”；feature extraction；target identification