卫星数字图像的智能处理研究

在借鉴成熟的数字图像特征分类模型基础上，通过特征抽取及采用神经网络方法对特征树训练，得到卫星数字图像的典型特征和在图像分析中的权重。提出利用数字图像的经验处理知识来实现卫星数字图像的自动处理，并建立分类数字图像集的经典处理案例，采用基于案例推理的方法，实现了卫星数字图像处理经验的复用和效率的提高。

Research on Intelligent Processing of Digital Satellite Imagery

Based on the mature classification model of the digital image, the neural network was used to train the feature tree and complete the feature extraction. With the weights of trained features, the typical features and their weights in the image analysis were obtained. Themethods of using the experience and knowledge of the digital images were proposed to realize the automated processing of the digital satellite imagery. Finally, by establishing the classical processing case of the classified image sets, the reuse of the processing experience and promotion of the processing efficiency were realized.