Laws纹理能量结合灰度共生矩阵的遥感影像面状地物提取  

Polygon Feature Extraction of Remote Sensing Image Based on Laws Texture Energy and Gray Level Co-occurrence Matrix

作　　者：李光[1] 姜春雪 刘争战 王立伟[1] 

机构地区：[1]天津市测绘院,天津300381 [2]天津市陆海测绘有限公司,天津300304

出　　处：《测绘与空间地理信息》2017年第7期179-181,185,共4页Geomatics & Spatial Information Technology

摘　　要：针对遥感影像中纹理较复杂的面状地物提取问题,提出一种Laws纹理能量结合灰度共生矩阵的遥感影像面状地物提取方法。该方法首先利用数学形态学膨胀、腐蚀对影像进行预处理;然后,分别对影像进行直方图均衡化,利用4个微观滤波算子滤波生成4幅微纹理影像;计算影像4个方向灰度共生矩阵及每个共生矩阵的4个纹理测度,生成4幅纹理影像;最后,结合8幅纹理影像进行区域生长跟踪轮廓。Aiming at extraction of texture more complex polygon feature in remote sensing image, a polygon feature extraction method of remote sensing image based on Laws texture energy and gray level co - occurrence matrix was presented. Firstly, corrosion and ex- pansion of mathematical morphology were used for image preprocessing. Then, image was processed by histogram equalization and four micro filter operator were used to generate four micro texture images. Four direction gray level co - occurrence matrix of image and four texture measure were calculated. Four texture images were generated. Finally, regional growth and trace contour were performed com- bining with eight texture images.

关 键 词：面状地物提取 Laws纹理能量 灰度共生矩阵 纹理影像 区域生长 

分 类 号：P283.8[天文地球—地图制图学与地理信息工程]