图像配准方法

说明：  通过超像素算法，读入图片确定分割数，得到最佳图像分割结果(Through the superpixel algorithm, read in the picture to determine the number of divisions to get the best image segmentation result)

用于abaqus建模内聚力单元建模，通过内聚力单元分析材料损伤(Used for abaqus modeling, the cohesion unit is used to analyze material damage)

说明：  IPC分类号的共现矩阵，技术融合，PYTHON代码(Co-occurrence matrix of IPC classification number)

python实现用逻辑回归识别和分类人工手写数字(Classifying MNIST digits using Logistic Regression)

算法思想来自于网上资源，先使用图像边缘和车牌颜色定位车牌，再识别字符。车牌定位在predict方法中，为说明清楚，完成代码和测试后，加了很多注释，请参看源码。车牌字符识别也在predict方法中，请参看源码中的注释，需要说明的是，车牌字符识别使用的算法是opencv的SVM， opencv的SVM使用代码来自于opencv附带的sample，StatModel类和SVM类都是sample中的代码。SVM训练使用的训练样本来自于github上的EasyPR的c++版本。由于训练样本有限，你测试时会发现，车牌字符识别，可能存在误差，尤其是第一个中文字符出现的误差概率较大。(Bits in the predict method, for clarity, add a lot of comments after completing the code and testing, see the source code. License plate character recognition is also included in the predict method. Please refer to the comments in the source code. It needs to be noted that the algorithm used for license plate character recognition is SVM of opencv. The training sample used for SVM training comes from the c++ version of EasyPR on github. Due to limited training samples, you may find errors in license plate character recognition when testing, especially the error probability of the first Chinese character is large.)

json文件转txt文件，批处理json文件转换成一个txt文件，需要json文件的哪部分只需要对函数做相应的修改即可，输入文件的路径就可以得到想要的结果，方便高效。

mqtt-python收发，里面有几个例子，对于topic的分类，mqtt的连接，收发等！(Mqtt transceiver in python!)

说明：  双目立体视觉标定以及立体匹配参考图片，可用于立体标定，图像校正，立体匹配算法验证(Binocular stereo vision calibration and stereo matching reference pictures can be used for stereo calibration, image correction and stereo matching algorithm verification)

音乐播放器，实现 wav 等多种音乐格式（至少 5 种）的打开、播放、暂停等功能。图形界面实现。文件内附不同格式的音乐

说明：  compression压缩感知算法用于图像处理。信号压缩(compressionCompressed sensing algorithm)