image-SVM

说明：  SIFT特征提取演算法------可用于两张影像之特征点匹配(SIFT feature extraction algorithm-can be used for feature point matching of two images)

对加载的图像进行灰度处理，简单处理matlab程序，效果不错，利于后面对图形进一步加工处理(Gray-scale processing of loaded images)

说明：  基于模糊C均值的快速点云的聚类分析代码，可用于模糊聚类(Clustering analysis code of fast point cloud based on fuzzy c-means can be used for fuzzy clustering)

说明：  使用pca方法对图像进行特征提取，对训练集的20个人的共一百张人脸进行训练，使用adaboost算法生成强分类器，可以对测试集的人脸图片进行识别，且识别率较高(The PCA method is used to extract the features of the image, and the training is carried out for a total of 100 faces of 20 people in the training set. The AdaBoost algorithm is used to generate a strong classifier, which can recognize the face images in the test set with a high recognition rate)

这个程序可以用于两个图像的拼接，通过寻找特征点完成拼接(This program can be used to stitch two images and complete the stitching by finding feature points.)

卡尔曼滤波是一种数据处理方法，它是一种线性最小方差无偏估计准则，基于系统 状态估计和当前观测，通过引入状态空间而获得的新的状态估计.本篇论文陈述了卡尔曼滤 波的基本思路和算法；并通过仿真，显示卡尔曼滤波的功能，以及如何用它来跟踪方向确定、速度恒定的飞行器。 (Kalman filter is a data processing method, which is a linear minimum variance unbiased estimation criteria, based on system state estimation and the current observations, obtained by introducing the state space of the new state estimate of the Kalman filter thesis statement the basic ideas and algorithms and through simulation, shows Kalman filtering functions, and how to use it to keep track of the direction determined constant speed aircraft.)

纹理特征提取佛挡杀佛规划环境和规范化多少倒萨打算颠三倒四的(Texture feature extraction block to kill the Buddha Buddha Planning, Environment and standardization of the number of Saddam' s intention to topsy-turvy)

根据超限像素平滑定义用 Matlab 自编写超限像素平滑法滤波(Super pixel smooth)

本代码用于实现.nc文件的IDL代码打开，并且可以实现，根据经纬度找到相应的像元值(This code is used to implement. Nc file IDL code open, and can be implemented, according to the longitude and latitude as yuan values can be found)

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.)