LRTsigma

说明：  harris检测 +ncc粗配+ransac去除错误点(harris detection. ncc coarse. ransac remove the error point)

国防科技大学博士周晓光论文-极化SAR图像分类方法研究(Research on polarization SAR image classification method by Dr. Zhou Xiaoguang of National University of Defense Technology)

显示BMP图片，并绘制直方图。程序无误，可运行。(Display BMP image, and draw the histogram. Program is correct, you can run.)

说明：  gui界面设计，进行图像的一般处理，如旋转，加噪，直方图均衡化等(gui interface design, the general image processing, such as rotation, noise, such as histogram equalization)

通过非监督分类中的K-均值聚类对Indian pines高光谱图像进行分类(Classification of hyperspectral images of pines Indian by K-means of non supervised classification)

说明：  MATLAB视频人体异常行为识别 本文设计了一款人体行为异常监控系统，主要适用人群是老年人，在摄像头固定的情况下，自动检测人体运动轨迹，并与提前设定好的行为库进行匹配，分析判断是否具有异常行为。(Human abnormal behavior recognition in matlab video In this paper, a monitoring system of human behavior abnormality is designed, which is mainly suitable for the elderly. When the camera is fixed, it can automatically detect the human movement track, and match with the behavior database set in advance to analyze and judge whether there is abnormal behavior.)

学习使用OpenCV，对输入的一段彩色视频，用OpenCV实现以下功能或要求: 1. 对输入视频的每一帧图像都用同一个阈值进行二值化； 2.在每帧二值化图像上叠加上含自己学号与姓名等信息的版权字幕； 3. 在处理的过程中，实时显示每帧图片处理之后的效果； 4.将所有这些二值化图像按视频原来顺序合成输出一个视频文件，按原输入视频播放速度的两倍合成； 5. 做成如下的命令行格式，xxx.exe 输入视频文件名 二值化阈值 输出视频文件名（例如 MyBiVideo.exe input.avi 50 output.avi ） 这里的二值化有两种选择，可以手动收入阈值，也可以直接用OTSU算法的较优阈值直接二值化。(Learn to use OpenCV, for some color video input, with the OpenCV following functions or requirements: 1. Each frame of the input video images are binarized using the same threshold value 2. superposed on the binarized image of each frame containing the copyright on the number and the name of their own learning subtitles other information 3. During processing, real-time display after processing each frame 4. All of these binarized image by synthesizing the video output of a video sequence of the original document, the original input composite video playback speed twice 5. made the following command line format, xxx.exe input video file name binary threshold value of the output video file name (eg MyBiVideo.exe input.avi 50 output.avi) Here there are two binary choice, you can manually income thresholds, you can directly use the optimum direct binarization threshold through OTSU algorithm.)

图像检测的算法，Faster R-CNN算法，先对整张图像进行卷积计算，然后通过感兴趣区域池化层(RoI Pooling Layer)将选择性搜索算法推荐出来的候选区域和卷积网络计算出的特征映射图进行融合，得到候选区域对应的特征矢量，这种共享卷积计算的操作极大地减少了卷积计算的次数。而且这些特征矢量的维度统一，方便后续的分类工作。通过感兴趣区域池化层处理卷积特征，并将得到的特征送往两个并行计算任务进行训练，分类和定位回归。通过这些方法和改进的框架，Fast R-CNN 用更短的训练和测试时长，取得了比 R-CNN 更好的效果(Faster R-CNN algorithm first convolutes the whole image, then fuses the candidate regions recommended by the selective search algorithm and the feature mapping maps calculated by the convolution network through the RoI Pooling Layer to get the corresponding feature vectors of the candidate regions, which greatly reduces the number of convolution calculations. Moreover, the dimension of these feature vectors is unified, which facilitates the subsequent classification work. The convolution feature is processed by the pooling layer of the region of interest, and the obtained feature is sent to two parallel computing tasks for training, classification and positioning regression. Through these methods and improved framework, Fast R-CNN uses shorter training and testing time and achieves better results than R-CNN.)

1.打开左右两张影像2.Harris算法提取同名点3.显示提取同名点（若要进行下一步匹配则不建议显示，这里只是用于查看是否提取出）4.相关系数匹配（只在找出的同名点内匹配）5.LSM最小二乘匹配（先进行粗匹配，再精匹配）(1 Open the left and right images 2.Harris algorithm to extract two tie points 3 shows the extraction of corresponding points (not recommended to proceed to the next match show, here is used to see if extracted) 4. Correlation coefficient matching (just looking match within a point of the same name) 5.LSM least squares matching (rough first match, then fine match))

HDR图像生成的算法，很实用和有效。利用多帧不同曝光时间的图像合成HDR的图像。(Algorithm of HDR image generation)