Application Of Remote Sensing Technology To Map Water Depth By Weakening The Influence Of Suspended Sediment

The means of remote sensing is the high-technology way to detect the water depth, which is known and accepted by people gradually. This means can not only get the large-scale and same-time underwater topography data, but also obtain the prophasic water depth data retrieved by historical satellitic image data, which provide the data to know variation of scouring and siltation and evolution of the riverway ulteriorly. Based on the important item of nation natural science fund "The effects and countermeasures of the water transfer projects etc. on the estuarine environment in the Yangtze River Estuary", the paper took the Yangtze River Estuary as a study object, the water depth mapping by means of remote sensing and spectral characters of suspended sediment were studied. The conclusions as follows:1. The water spectral reflectance of different suspended sediment concentration has the character of double peak, the first reflectance peak is in the wave bands of 550-720nm.the second lie in the 780-835nm.the percentage reflectance increases with suspended sediment concentration at all wavelengths, but the increasing amplitude is different. Results show that the sensitive wave bands of suspended sediment concentration are in the wave bands of R₁ (430-500 nm), R₂ (550- 570nm), R₃ (670-735nm) and R₄ (780-835nm).2. The percentage reflectance increases with a decrease of particle sizes at the same suspended sediment concentration, because the scatter surface area of suspended sediment increase with a decrease of particle sizes. The sensitive wave bands of sediment particle sizes are at the wavelengths of 500-700nm and about 800nm.3. The single wave band models, wave band ratio models and suspended sediment parameter models of suspended sediment concentration and its logarithm are constructed. Because the suspended sediment parameters have the suspended sediment information of many wave bands and partially decrease the effects by outside environment factors, the suspended sediment parameter models allow the suspended sediment concentration to be calculated at a better accuracy level than the single wave band models, wave band ratio models.4. The relationship between the spectral reflectance and the D₅₀^-1 of suspended sedimentis best, the universal expression between them is R = a + b · D₅₀^-1.5. Based on the analysis of the relationship between the spectral reflectance, the suspended sediment concentration, the particle sizes and wave bands, the relation model of them are constructed. Results show the model can express the relation among them well.6. Some water-depth retrieval models are constructed. Comparing the different models show that the line models are better than the non-linear models, multiple regress models are better than single factor regress models. Because the BP artificial neural network (BPANN) is a multitudinous parallel non-linear dynamic system, which exhibits a strong capability of non-linear mapping, the BPANN model can calculate the water depth value better than traditional statistic methods. The suspended sediment factor model can decrease the effects of suspended sediment on the water depth mapping by remote sensing, due to adding the suspended sediment factor, so it allows the water depth information to be retrieved at a relatively high accuracy level.