论文部分内容阅读
本文采用I~2S101数字图象处理系统对胶州湾幅CCT磁带数字图象信息进行了非监督和监督定量方法分类,分别获得了13类、9类及8类的地貌形态组合。结果表明,根据地物反射光谱的差异性,使用计算机来规定分类范畴的非监督分类方法(Non-Supervised learning)是成功的。 我们根据分类结果对胶州湾幅地貌形态特征进行了较理想的解译,特别是对湾内潮间带的水域及陆地分类清晰、准确。经实地考查,对解译成果作了验证。并且,通过计算
In this paper, I ~ 2S101 digital image processing system was used to classify the digital image information of Jiaozhou Bay CCT tapes by unsupervised and supervised methods. And then, 13, 9 and 8 topographic combinations were obtained respectively. The results show that non-supervised learning using a computer to classify categories is successful, based on the differences in reflectance spectra. Based on the classification results, we have conducted a better interpretation of the morphological characteristics of the Jiaozhou Bay, especially for the classification of the waters and territories in the intertidal zone of the Bay. After the field test, the result of the interpretation has been verified. And, by calculation