陳建國1，王晨輝2， 徐緒堪3 ，嵇慶才4

(1. 寧夏水利電力工程學校，寧夏 銀川 750006;2. 河海大學商學院，江蘇 常州 213022;3. 常州工業大數據挖掘與知識管理重點實驗室，江蘇 常州 213022;4. 鎮江新區城鄉建設局，江蘇 鎮江 212132)

　　摘 要： 在河湖“清四亂”專項活動中，全國各地對亂占、亂采、亂堆、亂建等河湖突出問題開展集中清理整治，針對人工排查工作中存在繁瑣、效率低下的問題，以水域遙感影像為研究對象，以提高違章建筑識別正確率和效率為目標，提出一種基于改進PCA和-Means的遙感影像變化檢測算法，該算法應用違章建筑的粗篩選和識別，對遙感影像進行幾何校正、轉換、特征空間提取、信息映射及構造圖像變化特征矩陣;通過改進-Means方法定位變化區域，給出水域違章建筑可能的區域，大大縮小人工檢測范圍。實驗結果表明，該方法可以快速、高精度地提取到用地變化信息，提高水域違章建筑識別效率和河湖監管能力。

　　關鍵詞：河湖“清四亂”;主成分分析;

　　-Means聚類;水域違章建筑識別

　　Research on recognition of illegal buildings in waters based on remote sensing images

　　CHEN Jianguo1, WANG Chenhui2, XU Xukan3, JI Qingcai4

　　(1.Ningxia Water Conservancy and Electric Power Engineering School, Yinchuan 750006, China;

　　2. Business School of Hohai University, Changzhou 213022, China;

　　3. Changzhou Key Laboratory of Industrial Big Data Mining and Knowledge Management, Changzhou 213022, China;

　　4. Urban and Rural Construction Bureau of Zhenjiang New District, Zhenjiang 212132, China)

　　Abstract: In the special activity of “clearing the four chaos” of rivers and lakes, all parts of China carry out centralized clean-up and remediation of prominent problems in rivers and lakes, such as indiscriminate occupation, indiscriminate mining, disorganization, and disorganized construction. Aiming at the problem of tedious and low efficiency in manual inspection work, to improve the accuracy and the efficiency of illegal buildings recognition, the article proposes a remote sensing image change detection algorithm based on improved PCA and k-Means . This algorithm is aimed to give a rough screening and recognition of illegal buildings. Remote sensing images perform geometric correction, conversion, feature space extraction, information mapping and construction of image change feature matrix. Then possible locations of illegal buildings are recognized by means of improved k-Means. It greatly reduce the scope of manual detection. Experiment shows that this method can quickly and accurately extract land use change information, improve the efficiency of recognizing illegal buildings in waters and the ability of river and lake supervision.

　　Keywords: rivers and lakes “clearing the four chaos”; principal component analysis; k-Means clustering; recognition of illegal buildings in waters