上海市黑臭水体光谱特征分析及遥感识别方法探究

Spectrum Characteristics Analysis of Black and Odorous Waters in Shanghai and Study of Remote Sensing Recognition

黑臭水体识别对全面开展黑臭水体整治及长效监管具有重要意义.为探究上海市黑臭水体的识别方法,以其2017年黑臭水体名录为研究对象,选取105条段河道样本,于2018年7—8月进行现场调查,开展水质监测、光谱测量、光谱分析、遥感识别及验证.结果表明:①依据上海市中小河道黑臭地方判定标准,样本中共有49个黑臭水体,56个一般水体.②将遥感反射率(R_(rs))作为光谱分类的主要依据,光谱数据预处理后,共得到4个主要反射峰(F)和3个主要吸收谷(G),根据波峰波谷位置特征,可将样本分为三类一般水体(GW)、三类黑臭水体(BOW).③R_(rs)平均值光谱曲线显示,BOW1、BOW3两类水体在550-700 nm处的光谱曲线与其他四类水体存在明显差异,更易识别和区分,而BOW2与GW2两类水体的光谱曲线形态差异较小,增加了识别难度.④根据拟合多光谱曲线,提出角度a1法和比值法组合成的角度-比值联合法,经识别验证,黑臭水体的综合识别率为95.92%,对一般水体的误判率为39.29%.⑤将归一化比值模型(BOI法)、黑臭水体斜率指数(SBWI法)、归一化黑臭水体指数(NDBWI法)三类已有识别方法用于105个样本识别,对黑臭水体识别效果均不佳,这可能是由于采样时间、地点不同,方法的判别阈值不适用等造成.研究显示,上海市黑臭水体中BOW1类和BOW3类水体与一般水体光谱特征有明显差异,角度-比值联合法对105个点位水样的总识别正确率达到77.14%,但该识别方法仅在上海市进行了探讨,还需在其他城市或地区进行验证.

The identification of black and odorous water is of great significance for comprehensive remediation of polluted water and long-term monitor of water quality.To develop the methods for identifying black and odorous water in Shanghai,105 rivers were selected from the list of black and odorous water of Shanghai City in 2017,and field investigation was conducted for water quality monitor,spectrum measurement,spectrum analysis,remote sensing identification and methods validation from July to August 2018.The results indicated that:(1)According to the local criteria of black and odorous water for small and medium rivers in Shanghai City,the rivers were divided into 49 black odorous water and 56 general water.(2)Using the reflectance of remote sensing(R_(rs))as the principle basis of spectrum classification,four main reflection peaks(F)and three main absorption valleys(G)were obtained after data preprocessing.According to the locations of the peaks and valleys,the samples were classified into 3 general water(marked as GW)and 3 of black odorous water(marked as BOW).(3)The figures of average values of R_(rs)showed that the spectrum of BOW1 and BOW3 were significantly different from the other four types of water samples at the wavelength of 550-700 nm,which were readily to be identified.However,the spectrum of BOW2 and GW2 exhibited minor differences,which increases the difficulty of identification.(4)Based on the simulation of the spectrum,the combined angle-ratio method that integrated 1 method with ratio method was proposed to identify the water samples.The comprehensive recognition rate of black and odorous water was 95.92%,while the misjudgment rate of general water was 39.29%.(5)Three existing achievement methods,BOI methods,SBWI methods,and NDBWI methods,were also used to identify the black and odorous water bodies.The poor recognition effect might be due to the different sampling time and locations,and the unapplicable model threshold.This study indicated that the spectrum characteristics of BOW1 and BOW3 in Shanghai City were significantly different from those of general water bodies.The total recognition rate of combined angle-ratio method for 105 water samples was as high as 77.14%.However,this recognition method is only used to study the water quality in Shanghai City and need to be further verified in other cities and regions.