(ok)NHLIScode

说明：  自组织神经网络SOM的算法，经过调试可以用(Self-organizing neural network SOM algorithm can be used after debugging)

圆形，孔型工件的尺寸测量，主要采用canny算子检测边缘，并进行亚像素细分，还有GUI，测量精度在0.1mm(Circular hole workpiece size measurement using canny operator edge detection and subpixel, as well as GUI, measurement accuracy of 0.1mm)

基于BP神经网络的人脸识别(java源码)

无源双基地雷达直达波脉冲丢失与相位突变影响分析Passive bistatic radar direct wave pulse loss and phase change impact analysis(Passive bistatic radar direct wave pulse loss and phase change impact analysis)

Zernike矩是一种具有尺度、移位和旋转不变性的正交不变矩，本设计的目的就是利用Zernike不变矩设计一种图像检索系统，该系统能够充分验证Zerinike矩的不变性及其在图像检索中的优良性能。具体内容包括： (1) 图像特征提取、统计特征提取； (2) Zernike不变矩及其应用方法； (3) 基于Zernike不变矩的图像检索系统。 (Zernike moments is a scale, shift and rotation invariant orthogonal invariant moments, the purpose of this design is the use of Zernike invariant moments design an image retrieval system, the system can fully verify Zerinike moments invariance and itsexcellent performance in image retrieval. Specific content includes: (1) image feature extraction, statistical feature extraction (2) Zernike invariant moments and its application method (3) based on Zernike Moment Invariant image retrieval system)

阈值重心法实现光带中心线提取，可用于结构光扫描的模型建立，处理二值化图像等。(Threshold barycenter method is used to extract the center line of optical band. It can be used to model structured light scanning and process binary images.)

采用gabor算法，灰度共生矩阵、灰度梯度共生矩阵提取图像纹理信息(Using Gabor algorithm, gray level co-occurrence matrix and gray level gradient co-occurrence matrix to extract image texture information)

说明：  在这个知识库中，图形信号处理的一些有趣的特性被表示出来。演示通过经典信号处理和图形信号处理对一维和二维欧几里得域信号应用低通滤波器，比较两种方法的结果。在此基础上，验证了图形信号处理的工作机制。此外，还描述了在非欧几里得域上滤波信号的示例。(In this repository, Some fascinating features of Graph Signal Processing were represented. Demos incudes applying a low-pass filter on both 1D and 2D euclidian domain signal by classical signal processing and also Graph signal processing to compare both results are the same. Within that way, we will validate the working mechanism of Graph Signal Processing. In addition to this, an example of filtering a signal on the non-euclidian domain was also represented.)

使用PCA和SVD进行数据去噪，利用数据主要的特征向量进行数据恢复和重建(Using PCA and SVD for data denoising)

说明：  取一个最优阈值把原图像分为前景色与背景色，两部分的类间方差越大，说明两部分差别越大，便能有效的分割图像。(Image threshold segmentation)