Bathymetry Retrieval From Hyperion Data In Shallow Water

Water depth is one of the most important habitat parameters of marine;it influences a lot of projects and activities in or on marine seriously,such as cultivation in ocean,the construction of roads and pipes on the ocean and so on.The water depth not only affects regional economic development and urban planning,but also affects national defense and planning.It usually needs to measure the water depth before engineering in or on marine.So bathymetry measurement is very important to city and nations.The traditional method of bathymetry measurement is to use ship-sounding techniques,which can provide high-precision data but is very costly and inefficient.This method needs a lot of people to complete the measurement and also needs some expensive measuring instruments.Because of the restriction of the method,operators have to measure the water depth point by point.Besides,it can't be complete in a bed weather or environment.With the development of remote sensing technology,a lot of researchers found that the data of remote sensing had a close relationship with the depth of shallow water;we can get the bathymetric of shallow water from these data.From the 60 s of twentieth century,the technologies of bathymetric measurement using remote sensing technology have been great development.The technologies usually include two methods,the active method and the passive method.The active method using microwave or laser radar to get the water depth directly,this method can get a more accurate water depth,but is also more expensive.The passive method is mainly using the images with the sun light to retrieve the bathymetry.Nowadays,the images of remote sensing,especially the hyperspectral images are easily to be get,more and more researchers have established the water depth inversion models based on hyperspectral images,these models can make use of the spectral information of hyperspectral images and the accuracy of the retrieved bathymetry is greatly improved.Based on the analysis of the Pearson correlation coefficient(CC)and the similarity coefficient(SC)of Hyperion images,we found that 1/ln(n * CC/SC)of Hyperion images had a close relationship with the water depths.Based on this research,we established the bathymetry retrieval model one.To analysis the validation of this model,two Hyperion data over the coast of Saipan and North Andaman Island were applied to demonstrate the proposed algorithm.The results showed that the model is effective and portable.During the analysis of the Pearson correlation coefficient and the similarity coefficient of the Hyperion images,we also found that(ln(nSC)-ln(nCC))/(ln(nSC)+ ln(nCC))had a close relationship with the water depths.Then we established the bathymetry retrieval model two.We used the same data used in the model one to analysis the validation of this algorithm.The high accuracy of the retrieved bathymetry indicated that the model two is effective and portable.In the end,the two bathymetry retrieval models that we established were compared and analyzed.The compared results showed that the two models can make good use of the hyperspectral spectral information,which can effectively reduce the influence of seabed sediment on bathymetry retrieve.The accuracy of the retrieved bathymetry using model one is slightly higher than the retrieved bathymetry using model two.