基于特征融合的区域生态环境动态监测研究  

Regional Environmental Monitoring Based on an Adaptive Fusion Method

作　　者：高歆[1] 徐永新[1] 吴国玺[1] 张京京 李长坡[1] 

机构地区：[1]许昌学院城乡规划与园林学院,河南许昌461000 [2]山西省环境规划院,山西太原030002

出　　处：《环境科学与技术》2015年第8期110-116,共7页Environmental Science & Technology

基　　金：河南省科技攻关计划项目(142102310201)

摘　　要：以1993、2000、2006和2013年4期遥感影像为主体数据源,选取山西省翼城县为实验区,利用遥感和GIS的技术手段对研究区生态环境进行动态监测和评价。为了更好地体现和突出图像中的纹理特征,满足矿区土地利用和生态环境状况调查解译精度的要求,在HIS小波融合的基础上,通过选取每个像元的特征向量并对其进行主成分分析,将获得的主成分作为高频系数融合规则权重来对原图像进行融合处理,并在此基础上将结果应用于生态环境要素提取和生态环境状况评价中。结果表明:该方法的标准差与信息熵最大,均值适中,在抑制光谱信息扭曲和提高图像清晰度等评价参数上优于其他方法;与1993年相比较,2013年高植被覆盖度区域减少了9 240.87 hm2,中度植被覆盖区域增加了14 036.44 hm2,中度、强度和极强土壤侵蚀面积分别增加了2 955.2、4 850.9和278.43 hm2;实验区生态环境质量指数由37.5上升到40.7,2013年又下降到了33.1,表明了20年间生态环境质量呈现出先改善后恶化的局面。With Yicheng County located in southern Shanxi Province as study area, and its four remote sensing images during the period from 1993 to 2013 as data source, dynamic monitoring and evaluation of ecological environment were done based on RS and GIS technology. In order to enhance textile information and interpretation accuracy, a method, which is in terms of fusion rule using principal component weighted for high frequency coefficient from feature vector of each pixel, was proposes and used for extraction of ecological environment elements and estimation of environment. Results indicated that the method can effectively inhibit spectral distortion and improve image definition with larger information entropy and standard deviation than other methods. High vegetation coverage area decreased by 9 240.87 hectare, while medium vegetation coverage area increased by 14 036.44 hectare, in addition, medium, serious and very serious soil erosion areas increased by 2 955.2, 4 850.9 and 278.43 hectare respectively. Environmental quality index calculated from five factors increased from 37.5 to 40.7, and then decreased to 33.1, which indicated that it got better at first, and afterward got worse in 20 years.

关 键 词：小波变换 HIS-WT融合 生态环境 遥感 评价 

分 类 号：X83[环境科学与工程—环境工程]