基于CCD数据的太湖蓝藻水华监测算法研究

发布时间：2018-11-12 20:47

【摘要】：近些年来,随着我国工、农业的迅速发展,大量含有氮、磷等元素的工业、农业污水和生活废水被排入到河流与湖泊中,使得我国湖泊出现了严重的富营养化现象,尤其是长江中下游一带的湖泊,例如:太湖、巢湖等。每年,夏秋季节这些湖泊都会暴发严重的蓝藻水华。蓝藻水华直接影响着人类的健康、经济的发展以及生态的平衡。本文主要利用我国国产的环境卫星CCD数据,通过自动控制散点回归的方法(ASCR)对不同影像的CCD数据进行相对辐射校正,再将归一化植被指数(NDVI)和像元生长算法(APA)相结合,提出了一种可业务化运行的蓝藻水华高精度提取算法,并将该算法应用在新发射的高分一号卫星和美国陆地Landsat卫星上。通过本文的研究,可得出以下结论:(1)通过自动控制散点回归方法进行相对辐射校正后的CCD影像,校正结果较为理想,使得相对稳定的同名地物在不同时相影像上辐射值一致,以此通过不同时相影像上辐射值的差异来进行太湖蓝藻水华动态变化的监测。(2)利用NDVI指数初步提取每景影像上太湖区域的蓝藻水华,接着利用坡度分析法确每一景影像的阈值,再利用统计分析的方法,确定一个统一的蓝藻水华提取阈值,解决了以往一景影像一个阈值,难以大规模批处理的难题。(3)通过像元生长算法对像元进行线性分解,提取精度可以达到亚像元级别,更精确的统计了太湖蓝藻水华的面积和分布情况。(4)对2009-2014年下半年太湖蓝藻水华暴发情况进行长时间序列的连续监测,发现2013-2014年太湖蓝藻水华较以往相比暴发面积减少,规模较小,水质得到了控制和改善。研究也表明,该算法对蓝藻水华识别能力强,自动化程度和水华提取精度高,可作为业务化运行的算法。(5)对比我国新发射的高分一号卫星CCD传感器,以及美国的Landsat卫星传感器,发现它们与环境卫星CCD传感器的相关性较高。由此,将本算法应用在高分一号卫星数据,以及美国Landsat系列数据上,基于环境卫星CCD数据对其它影像数据进行相对辐射校正,最终实现了利用多卫星平台进行更长时间序列和更高时间分辨率的太湖蓝藻水华的实时动态监测。
[Abstract]:In recent years, with the rapid development of industry and agriculture in China, a large number of industrial, agricultural sewage and domestic wastewater containing nitrogen, phosphorus and other elements have been discharged into rivers and lakes, resulting in serious eutrophication of lakes in China. Especially the middle and lower reaches of the Yangtze River lakes, such as: Taihu Lake, Chaohu Lake. Every year, in summer and autumn, these lakes will break out serious cyanobacteria Shui Hua. Cyanobacteria Shui Hua directly affects human health, economic development and ecological balance. In this paper, the relative radiation correction of CCD data of different images is carried out by means of automatic control of scatter point regression (ASCR), which is based on the CCD data of the environment satellite made in China. Combining normalized vegetation index (NDVI) with pixel growth algorithm (APA), a high precision extraction algorithm for cyanobacteria Shui Hua is proposed. The algorithm is applied to the newly launched Gaofen 1 satellite and the American land Landsat satellite. Through the research of this paper, the following conclusions can be drawn: (1) the CCD image after relative radiation correction is obtained by automatic control scatter point regression method, and the correction result is satisfactory. Making the radiation values of relatively stable objects of the same name consistent in different phase images, Therefore, the dynamic changes of cyanobacteria Shui Hua in Taihu Lake were monitored by the difference of radiation values in different phase images. (2) the blue algae Shui Hua in Taihu Lake region was preliminarily extracted by using NDVI index. Then the threshold value of each scene image is ascertained by slope analysis method, and then a unified blue algae Shui Hua extraction threshold is determined by using statistical analysis method, which solves the problem of one threshold value of a scene image in the past. It is difficult to solve the problem of large-scale batch processing. (3) the pixel growth algorithm is used to decompose the pixel linearly, and the extraction precision can reach sub-pixel level. More accurate statistics on the area and distribution of cyanobacteria Shui Hua in Taihu Lake were made. (4) continuous monitoring of a long-term series of outbreaks of cyanobacteria from Lake Taihu in the second half of 2009-2014 was carried out. It is found that the outbreak area of cyanobacteria Shui Hua in Taihu Lake in 2013-2014 is smaller than that in the past, and the water quality has been controlled and improved. The results also show that the algorithm has strong recognition ability to cyanobacteria Shui Hua, high degree of automation and extraction accuracy of Shui Hua, and can be used as an algorithm for operational operation. (5) comparing with the new CCD sensor launched by China's high score 1 satellite, And the Landsat sensors of the United States, they are found to be highly correlated with the environmental satellite CCD sensors. Therefore, the algorithm is applied to the Gaofen 1 satellite data and the Landsat series data of the United States. Based on the environmental satellite CCD data, the relative radiometric correction of other image data is carried out. Finally, real-time dynamic monitoring of cyanobacteria Shui Hua in Taihu Lake with longer time series and higher time resolution is realized by using multi-satellite platform.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：X87

【相似文献】