Threedimension

说明：  提取图像中的模式噪声，是先使用滤波器后再进行提取的(To extract the pattern noise in an image, a filter is used first and then the noise is extracted)

自己实现的BM3D图像去噪算法。它首先把图像分成一定大小的块，根据图像块之间的相似性，把具有相似结构的二维图像块组合在一起形成三维数组，然后用联合滤波的方法对这些三维数组进行处理，最后，通过逆变换，把处理后的结果返回到原图像中，从而得到去噪后的图像。(Image Denosing by Sparse 3-D Transform-Domain Collaborative Filtering.)

副高五个指数的计算，可根据需要修改，数据需要提取，注意修改区域(the index of subtropical high)

visualC 数字图像模式识别与技术详解，书籍源码部分。(visual C digital image pattern recognition and techniques detailed books source parts.)

完成水果的尺寸形状表面缺陷的判别，BP神经网络的训练判别表面缺陷(Fruit size, shape and surface defects are identified. BP neural network is used to identify surface defects.)

利用opencv基于灰度投影算法进行图像平移补偿(Use opencv gray projection algorithm for image translational compensation)

说明：  Delphi利用OpenCV实现的视频录像单元(Implementation of capture camera to avi by OpenCV in Delphi.)

基于harris兴趣点的图像检索程序。先特征提取，进而进行特征匹配，最后还有结果的返回。(Harris points of interest based on image retrieval process. First feature extraction, feature matching and then carried out and, finally, the results returned.)

基于mmdetection框架的深度学习目标检测框架(deep learning object detection based on mmdetection)

该算法是通过对图像使用色彩滤波矩阵来检测图像是否进行了篡改(The algorithm detects whether the image is tampered with by using the color filter matrix.)