zhangwen

基于Curvelet变换的样本块图像修复算法提高现有样本块修复算法性能。首先利用Curvelet 变换估计待修复图像的4方向特征．然后利用颜色信息与方向信息共同衡量样本块间的相似度，在此基础上构造颜色-方向结构稀疏度函数．(Based on the sample block transform image restoration algorithm Curvelet enhance existing sample block repair algorithm performance. Firstly Curvelet transform estimated to be repaired 4 directional characteristics of the image. Then use the color information and direction information together measure the similarity between the sample block, constructed on the basis of color- direction sparsity structure function.)

有关图像的目标识别："给出一种基于特征分类辨识的合成孔径雷达图像目标检测方法#用恒虚警和扩展分形方法对3&E图像进行目 标检测后用面积和峰值能量比算子辨识目标和背景杂波!去除一部分虚警!用小波域主成分分析对每个检测窗口内的图 像提取特征向量!用支持向量机对提取得到的特征向量进行分类!辨识目标和背景杂波!完成目标检测#使用&K?3数 据对该方法进行验证和分析!实验结果表明!经过特征分类辨识后!在检测率不变的情况下!虚警数目显著降低# (Related to the image target recognition)

这个压缩包中包含了巴特沃斯滤波算法的高通和低通两款代码，同时还有几篇巴特沃斯算法的文章，记录了关于巴特沃斯算法的一些应用领域。(This compressed package contains the code of two of the high-pass and low-pass Butterworth filter algorithm, at the same time, there are few Butterworth algorithm articles recorded Butterworth algorithm applications.)

图像的阈值分割，基于matlab的一维非广延熵Tsallis熵算法的灰度图像分割(Image segmentation, matlab-based one-dimensional non-extended entropy Tsallis entropy algorithm gray-scale image segmentation)

增量直线拟合(incremental line fitting)算法描述多个物体(Incremental line fitting (incremental line fitting) algorithm for multiple objects described)

说明：  对图像畸变进行校正,对有些图像由于近小远大照成的畸变进行校正，采用了opencv(To correct image distortion , It s important to the machine vision)

用水线阈值法分割图像：通过对图像进行高帽、低帽变换，并将得到的结果进行相减，可以增强图像，然后利用分水岭算法就行分割！(Water line thresholding segmentation image: The image of flattery, low cap transformation, and will be the results of subtraction, it can enhance image, and then use the watershed segmentation algorithm on the line!)

双目匹配计算视差，生成深度图。（前提是用自己标定好的摄像头拍的图片，导入内外参数然后进行立体匹配）(Binocular matching calculates parallax and generates depth map.(The premise is to take a picture with your own calibrated camera, import internal and external parameters, and then stereo matching))

本文首先对数字全息技术的记录和再现的基本原理、数值重建算法、频谱滤波方法，数字全息图的记录条件、再现光场的分离条件等基本问题进行了分析讨论(Firstly, the basic principle of the recording and reproducing of the digital holography, numerical reconstruction algorithm, the spectral filtering method, the digital hologram recording conditions, reproducing conditions of separation of the light field, and other basic issues discussed)

gist特征用于商品图像检索，其为全局特征，检索效果非常好。(GIST feature is used to realize image retrival it performs very well.)