SAR强度分布拟合

亚像素边缘检测，基于zernike 灰度矩，实现精度可达0.05个pixel(sub-pixel edge detect)

2维的基于互信息的图像配准算法，使用powell法作为优化算法(2-dimensional mutual information based image registration algorithms, the use of law as the optimization algorithm powell)

高斯混合模型的最大期望迭代求解算法,可用于图像区域灰度分布估计(Expectation maximazation(EM) for Gaussian mixture model(GMM) )

python写的图像相似度算法。比较图像相似度，选出最相思的图片(python write the image similarity algorithm. Compare image similarity, image to select the most Acacia)

图像压缩相关代码，基于小波域的矢量量化算法实现图像编码功能(Image compression code image coding, wavelet-based vector quantization algorithm)

小波分析在图像处理中有非常重要的应用，包括图像压缩，图像去噪，图像融合，图像分解，图像增强等。文中给出了详细的程序范例，用MATLAB实现了基于小波变换的图像处理。(Wavelet analysis in image processing has very important applications, including image compression, image denoising, image fusion, image decomposition, image enhancement. The paper gives detailed procedures for example, implemented with the MATLAB image processing based on wavelet transform.)

改进lee滤波源码，里面有im是原图 nim是噪声图像

说明：  k-svd算法用于稀疏表示的图像去噪，字典学习算法，去噪效果良好(k-svd algorithm for image denoising of sparse representation)

一个自由、开源的图形库--FreeImage3.92最新版的中文参考(支持位图构件―例如调色板和数据位―的便易访问；将位图从一种位深度转换到另一种位深度；当有多幅位图页―例如TIFF―时访问位图的不同页；基本的位图处理，如旋转、翻转（flipping）和重采样，或点操作，如亮度和对比度调整；Alpha混合与合成（Compositing and blending）；FreeImage不支持：高级图象处理操作，如回卷（Convolution）和变换（Transform）；位图绘制； 矢量图形。)(a free, revenue graphics library-- the latest version FreeImage3 .92 reference to the Chinese (support bitmap elements- such as transfer shade and data spaces- the easy access; will be from a bitmap-depth conversion to another kind of depth; When a number of bitmap page- such as TIFF- Bitmap to visit the different pages; basic bitmap processing, such as rotation, Upset (flipping) and resampling, or point operations, such as brightness and contrast adjustments; Alpha mixed with the synthetic (Compositing and blending); FreeImage not support : Advanced image processing operation, If the response (Convolution) and transform (Transform); bitmap rendering; Vector graphics.))

通过使用块匹配法实现运动矢量的检测，块大小和搜索区域大小可自定(Achieved through the use of block matching motion vector detection, block size and search area size can be customized)