Environmental Evaluation of Seto Inland Sea by Remote Sensing

近年来，酸雨问题、温室效应、臭氧层破坏等各种环境问题成为人们关注的焦点，这些问题严重影响着我们未来的生活。在这项研究中，我们考虑了利用遥感数据分析的方法对泽东内海进行环境评价。特别是，从全球环境角度来看，温室效应问题已经变得重要起来。对于这样的问题，遥感技术适用于获取大范围、大范围的环境状况。但是，为了将环境状况作为二维图像来获取，我们需要地表的环境数据，这些数据被称为宏观真实数据。为此，我们于1990年收集了分布在中部或四国地区的后藤内海水污染数据，并建立了地面实况数据数据库。在此基础上，利用遥感分析和地面实况数据，对1991年的泽东内海陆地卫星TM数据进行了分析。最后，在1992年，我们讨论了所获得的图像与地面真实数据之间的误差。此外，为了获得比传统的贝叶斯分类方法更精确的分类结果，我们还引入了一种新的基于神经网络的遥感分类方法。

Recently, various environmental problems such as acid rain problem,green house effect, and ozonn layer damage have been focused on sice those problems are serious for our lives in the future. In this study, we have considered environmental evaluation in Seto Inland Sea by remote sensing data analysis. Especially, a green house effect problem has become important from global environmental viewpoints. For such a problem, the remote sensing technique is suitable for getting an environmental situation in a wide, range. But in order to get the environmental situation as a two-dimensional image, we need the environmental data on the surface, which are called grand truth data. Therefore, in 1990 we have gathered the water pollution data in Seto Inland Sea which have been distributed in Chugoku or Shikoku areas and constructed the data base about the ground truth data. Then in 1991 we have analysed the LANDSAT TM data in Seto Inland Sea by using the remote sensing analysis and ground truth data. Finally, in 1992 we have discussed the error between the obtained image and the ground truth data. Furthermore, we have introduced a new remote sensing method based on neural networks in order to obtain more precise classification results compared with a conventional Bayesian method.