中文摘要：

水体总悬浮物浓度影响水体的光场分布，进而影响初级生产力。2012年7月14—15日，在松花江哈尔滨段选取11个断面共33个样点，利用野外高光谱仪（ASD Field Spec Pro）进行反射光谱测量和同步水样采集分析，通过研究总悬浮物质量浓度与其高光谱反射率之间的相关关系，采用单波段反射率、反射率比值法和一阶微分法分别建立了总悬浮物浓度的高光谱估测模型。结果表明，3种方法的决定系数都在0.90左右，均可以用于总悬浮物的定量遥感监测，其中一阶微分模型为反演的最佳模型。

英文摘要：

In accordance with the high concentration features of the total suspended solid concentration （TSS）, this paper has done an analysis of the Pearson coefficients between TSS and field spectra. As a research result, the total suspended solids concentration we have adopted is founded based on the hyperspectra data provided by the field spectral statistics in July, 2012. What we want to expound in this paper is to make clear a single or a few waveband combinations which can be used for establishing an estimation model of the total suspended solids concentration. It is just for this purpose that we have resorted to the hyperspectral remote sensing bands that can make the hyperspectral remote sensing most sensitive to the total suspended solids concentration. Besides, we have to make clear the correlation between the TSS and high spectral reflectance though the correlation coefficient may not be very exact. What is more, it is necessary to normalize the field spectral and make the first-order differential processing. And, for this need, when we make an analysis of the correlation between TSS and the normalized spectral reflectance and one-order-derivative, the software Excel can help us to find the high Pearson coefficient. It is just by using the above said three methods of the single band, the spectra ratio of the two bands and the one-order-derivative that can contribute to the establishment of an estimation model of the total suspended solids concentration. In the single band correlation coefficient, the reflectance of 773 nm is the highest one, from which we can deduce the estimation model （ y =12.615e 3.0539 x , R 2 =0.91） between the TSS and the reflectance of 773 nm. In spectra ratio of the two bands, the correlation coefficient of R 773 / R 657 should be the optimal one,that can help us to establish an estimation model （ y =13.001e 3.733 x , R 2 =0.92）. In the one-order-derivative, the correlation coefficient of 711 nm is the highest and may conduce to the establishment of an estimation model （ y