Research For Abstraction Of The Information Of Ocean Water Quality With Remote Sensing

Near twenty years, along with the development of economic, plenty of industrial waste water and sewage was drained into the sea and made the seawater greatly polluted, which led to sequential red tides and other environmental disasters, and the economy was greatly damaged. So we must pay much attention to ocean environment for the case of the balance of oceanic environment and the sustainable utilization of oceanic resources. We must supervise and research the degree and quality of the pollution of the whole ocean, in order to control and reduce the pollution. The conventional survey method has many limitations. From these data we can only get few discrete point values in rapidly and complicatedly with the current especially in coast areas. So the data acquired with conventional method can't satisfy concentration research. The application of remote sensing technology makes it possible to get water image in large scale and quickly. And the images contain a great deal useful information on seawater. Its quality of instantly synchronization is good, and its cycle of capturing data is short. Now it gradually becomes a very important method to supervise ocean water quality.In this paper , spectral mixing analysis was used on Band 1, 2, 3 of TM image to estimate the suspended sediment concentration, a model describes the relationship between high suspended sediment water and the concentration was acquired, and the result was compared with that of negative exponential model. Empirical model, neural network model and spectral mixing analysis were used on SeaWiFS image to estimate the chlorophyll concentration is Taiwan Strait, and the rule of chlorophyll distribution was concluded. Then three methods were evaluated by their statistical criteria, graphical criteria and the image calculated results. Spectral mixing analysis was used for retrieval of multiconstituents in seawater. The concentration of suspended sediment, chlorophyll and yellow substance at the same time, then it will be more accurate for it estimates one continent in water while removes the effects of others. It was concluded that the spectral mixing analysis is a good method especially in areas lack of much measured data for it can make full use of multi-band of images, needs less samples, and can be used to multicondtituents water color remote sensing easily.