colorimagerestoration

针对CMOS 图像传感器输出的Bayer 格式的图像数据, 利用TI 公司的TMS320DM642 数字信号处理器的专用视频端口,可以实现图像捕获、数据传输的无缝连接的特点, 构建了以DM642 为核心的图像实时处理系统。在DM642 中采用两种通用方法(双线性插值、边缘检测) 和文中提出的信号相关的方法, 将从Bayer 公司的CMOS 图像传感器所捕获的8 位Bayer 格式图像数据, 转换为24 位RGB 格式彩色图像数据, 并通过PSNR(信噪比) 和NCD(归一化彩色差异规范) 两种评估方法进行检测。从实验结果可以看出文中提出的方法相对于通用方法提高了图像的信噪比, 锐化图像的边缘, 减少了图像高频信息的错误像素的数量, 提高视觉质量。(Bayer TMS320DM642)

This is the basic code implementing jpeg2000 in matlab

求图像的二维直方图，产生共生矩阵，可以比较像素点分布的是否均匀。(Two dimensional histogram for image, the co-occurrence matrix, can compare the pixel distribution is uniform.)

说明：  可以对光谱数据进行预处理，除了偏最小二乘法之外，也有一些其他方法(Besides the partial least square method, there are also some other methods)

In the use of MATLAB to complete the wavelet transform and threshold to the process of image compression, very good

压缩感知中利用增广拉格朗日方程解最小稀疏正则化的恢复算法(DAL solves the dual problem of (1) via the augmented Lagrangian method (see Bertsekas 82). It uses the analytic expression (and its derivatives) of the following soft-thresholding operation， which can be computed for L1 and grouped L1 (and many other) sparsity inducing regularizers. If you are interested in our algorithm please find more details in our technical report or in my talk at Optimization for Machine Learning Workshop (NIPS 2009).)

非局部平均，用来去除灰度图里面的噪声，效果明显(Probability Non-Local Means)

用Visual C++实现基于数学形态学的数字图像处理,包括二值图像的腐蚀,膨胀,开,闭,细化及中轴变换等运算.(Visual C++ using mathematical morphology-based digital image processing, including binary image erosion and dilation, opening, closing, thinning, and the central axis transform operations.)

说明：  小波图像融合的6种方法，最大值、局部能量、平均值等等(There are six methods of wavelet image fusion, maximum value, local energy, average value and so on)

对植物图像进行图像分割，区域标记，区域提取(Images of plant image segmentation, region markers, regional extraction)