

摘 "要: 以農業用地與人工建筑的分割為研究目的,提出一種可在衛星或航拍遙感圖像中自動分割農用田地與人工建筑的圖像處理算法。該算法使用圖像區域合成、濾波、灰度統計、背景分割等處理技術,可實現大面積農用田地中人工建筑及道路的分割,并給出面積的估算。在此給出了解決思路,并給出試驗效果。該技術對提高農地統計自動化水平、農業用地侵占普查效率有一定工程意義。
關鍵詞: 耕地; 人工建筑; 閾值; 圖像分割
中圖分類號: TN964?34 " " " " " " " " " 文獻標識碼: A " " " " " " " " " " " " " "文章編號: 1004?373X(2015)06?0105?03
Research on automatic segmentation of agricultural land and artificial buildings in remote sensing image
WANG Min?yan1, ZHAO Kun2
(1. Commission for Discipline Inspection and Supervision Bureau of Jining, Jining 272045, China;
2. The 27th Research Institute of China Electronics Technology Group Corporation, Zhengzhou 450047, China)
Abstract: An image processing algorithm which can automatically divide up agricultural fields and artificial buildings in remote sensing images from satellite or aerial photo is proposed. With the technologies of image area synthesis, filtering, histogram statistics and background segmentation, the artificial buildings and roads in large?scale agricultural lands can be segmented by the algorithm, and the area calculation can be realized. Some solutions and test results are given in this paper. It is significant for this algorithm to improve the automation statistic level of agricultural land.
Keywords: agricultural field; artificial building; threshold; image segmentation
0 "引 "言
作為人口大國,農林業對我國家安全和社會穩定有重要意義,隨著經濟的高速發展,鄉鎮及城市建設對土地需求日趨增長,合理有效地使用土地資源,監控土地實際使用情況是國家研究的重要課題。隨著現代航天技術、航空技術、探測器技術高速發展,國家投入大量衛星、飛機及無人機進行耕地普查,新技術的廣泛應用給農業用地統計、普查、災害預警帶來革命性發展,提供了豐富的數據資源。但社會各界對農業統計信息的需求越來越大,對統計質量要求越來越高,按照傳統的方法,僅僅依靠手工作業是遠遠不能適應實際海量數據篩選需要的。本文提出一種可在衛星或航拍遙感圖像中自動分割農用田地與人工建筑的圖像處理算法,該算法使用圖像區域合成、濾波、灰度統計、背景分割等處理技術,可實現農用田地中疑似建筑及道路進行分割,并給出面積的估算。……