基于灰度共生矩阵的未利用地疑似污染遥感识别

Monitoring Suspected Pollution on Unutilized Land Using Gray-level Co-occurrence Matrices

为了加强大面积范围内未利用地监管,提出通过遥感技术识别存在潜在污染的未利用地.以甘肃省北部地区为研究区,首先,基于Landsat卫星数据进行土地利用/覆被类型遥感解译,确定该区域未利用土地范围.其次,对图像进行主成分分析,将第一主分量作为灰度共生矩阵的数据源,选用能量、熵、惯性矩、相关作为特征量,同时结合对应图像的灰度变化绝对值提取变化较大的区域.最后,通过对比2010年和2015年Landsat遥感图像的特征量变化情况,提取有明显纹理或灰度变化区域,结合Google Earth高分辨率影像与包含工矿企业位置信息的感兴趣点(point of interest,POI)数据,得到2010—2015年此区域土壤疑似污染点40处,总面积约为10 km2.对其中21处结果进行实地调查验证,其中有19处疑似污染点被证实,识别精度约为90%.提出的基于灰度共生矩阵方法识别未利用地疑似污染的方法,较传统人工解译方法,能够显著节省人力、物力,提高监测效率,并且具有较好的精度.

To strengthen the supervision of unutilized land in large areas,this paper identified the unutilized land of potential pollution through remote sensing technology. Taking the northern part of Gansu Province as the research area,a remote sensing interpretation of the land use and the cover type based on Landsat satellite data was first conducted to determine the scope of the unused land in the area.Second,the principal component analysis（ PCA） of the image was carried out,and the first principal component was used as the data source of the gray level co-occurrence matrix. The energy,entropy,moment of inertia and correlation were selected as the feature quantities. The absolute value of the gray value change of the corresponding image was combined to extract the greatly-changed area. Finally,by comparing the feature quantity changes of Landsat remote sensing images in 2010 and 2015,areas with obvious texture or grayscale changes were extracted. Combining with the high-resolution images of Google Earth and point of interest（ POI） data containing industrial and mining enterprise location information,it was concluded that there were 40 suspected soil pollution sites in this area from 2010 to 2015,with a total area of about 10 square kilometers. Twenty-one of the results were conducted field-test,and 19 suspected contamination sites were confirmed, with an accuracy of approximately 90%. Compared with the traditional manual interpretation method,the method based on the gray level co-occurrence matrix method to identify the suspected pollution in unused land is more effective. It can save manpower and material resources significantly,improve the monitoring efficiency and have better precision.