Digital-Image-Processing-By-Gonzalez-2nd-Edition-

边缘检测的绝好套餐，可以拿来做博士 论文了(Edge Detection of excellent packages, can be used to do a doctoral dissertation)

图形学中的透视投影算法 在屏幕中心绘制立方体的透视投影 可以绘制一点透视图 二点透视图 三点透视图(Graphics in the perspective projection algorithm for drawing a cube in the center of the screen you can draw a little perspective projection perspective two point perspective three-point perspective)

说明：  my_noise文件是已知信号方差，加入信号噪声，可以加入高斯噪声，淑盐噪声,等其他噪声十分有用 其他两个函数时加入噪声信噪比SNR,但是有一点,这种方法并不准确,不能保证加入噪声信噪比为SNR,实际测试中可以看出,但是可以加入相同的信噪比的噪声(my_noise document is known to signal variance, by adding signal noise, Gaussian noise can be added, suk salt noise, noise, and other very useful function when the other two noise signal to noise ratio SNR, but there is one point, this method is not accurate , can not be guaranteed by adding the noise signal to noise ratio for the SNR, the actual test can be seen, but can join the same noise signal to noise ratio)

基于多阈值的图像分割()

图像质量评价中全参考方法IWSSIM的代码，是经典方法SSIM的升级(The code of IWSSIM, a full reference method in image quality evaluation, is an upgrade of SSIM, a classical method)

这是DCT的改进算法，可以提高DCT算法的执行速度。(This is the DCT of the improved algorithm can improve the execution speed DCT algorithms.)

说明：  使用中值，均值，小波，DCT，PCA五种方法实现对图像的去噪处理。(Five methods, median, mean, wavelet, DCT and PCA, are used to denoise the image.)

how to use win-logon

非下采样轮廓波变换，NSCT变换，主要用于图像去噪(Non - down sampling contourlet transform and NSCT transform are mainly used for image denoising.)

说明：  连通监控视频的 火灾检测，颜色特征，纹理特征等(fire detection Description for each code - extract_red.m, objectlabel.m : Ground Truth labeling automation code generated from Apps - FireDetection.m : MATLAB class to import custom algorithm to Image Labeler - yolov2_detect.m : Entry function for GPU Code generation - GroundTruthTest.mlx : Test code for ground truth labeling - yoloFire_Train.mlx : Training code for YOLOv2 network - Live_Webcam_YOLOv2.mlx : Live webcam demonstration code for YOLOv2 network)