论文部分内容阅读
所谓图像增强处理,系指能提高图像数据目视质量的各种处理,大多数增强是变动原始数值扩大图像反差,或以彩色编码提高图像目视分辨效果。在整个处理过程中,计算机算法没有“训练学习”的机会,即未应用“已知信息”对算法进行“监督”。本文笔者拟用图像监督分类法的基本设想,给出线性组合模型,选择训练场,在一定统计意义下用训练样本估计线性组合模型的权系数,得到信息统计特征增强的综合图像.为此,需将已知信息揉和到图像信息增强技术中去,有目的地增强和提取感兴趣的图像特征信息。这里对方法和效果作一简要介绍。
The so-called image enhancement processing, refers to various processes that can improve the visual quality of image data, most of the enhancement is to change the original value to enlarge the image contrast, or color coding to improve the image visual resolution. Throughout the process, computer algorithms do not have the “training learning” chance of not using “known information” to “monitor” the algorithm. In this paper, the author intends to use the basic idea of the image supervised taxonomy to give a linear combination model, select the training field, and use the training samples to estimate the weight coefficients of the linear combination model in a certain statistical sense to obtain a comprehensive image with enhanced information statistical characteristics.Therefore, The known information needs to be kneaded into the image information enhancement technology to purposefully enhance and extract the image feature information of interest. Here’s a brief introduction of methods and effects.