yaogan

说明：  有数字1和2的图片各四十张，利用mat lab读取图片并将其二值化，得到只由0和1表示的矩阵，根据数字的不同特征，选择两个特征值，用来对分类器进行训练，最后用剩下的5个1和5个2来进行测试，看分类器是否可以正确的对数据进行分类。(Figures 1 and 2 pictures of the 40, using mat lab picture and secondly, the value of reading, and received only by 0 and 1, said matrix, according to the number of different features, select the two characteristic values, used to classifier training, and finally with the remaining 5 1 and 5 2 to carry out tests to see whether the classifier can correctly classify the data.)

利用SIFT检测两图片的相同点，并获取两幅图像的视差。(SIFT detection using the same picture two points, and to obtain two images of the parallax.)

集运小波变换的图像去噪，本文采用的是一种改进了的软阈值法去噪，去噪效果更好(Image Denoising Based on Wavelet changing)

超声图像处理软件dream，开源代码，用于仿真超声图像(Ultrasound image processing software dream, open-source code for the simulation of ultrasound images)

Metaface Learning for Sparse Fisher Discrimination Dictionary Learning for Sparse Representation

Lee的经典的多分辨率B样条拟合方法（multi-level B-spline fitting）的VC源代码。广泛应用于图形图像处理，CAD等领域。(Lee s classic multi-resolution B-spline fitting method (multi-level B-spline fitting) of the VC source code. Widely used in graphic image processing, CAD and other fields.)

复杂场景中的红色交通标志牌检测与识别，结合颜色与圆形的标志牌定位，然后提取特征，最后用模板匹配的方法实现识别，内附图片和程序说明，对初学者有帮助(Red detection and recognition of traffic signs in complex scenes, the combination of colors and circular signs positioning, and feature extraction, the last with template matching methods to achieve recognition, included pictures and a description of the procedures for beginners)

一些matla实现的图像处理算法，有emd,和BEMD分解的程序，及图像去噪等算法。(Some matla realize image processing algorithm, emd, and BEMD decomposition process, and image denoising algorithm.)

说明：  图像融合是图像处理的一个研究热点，这里提供一种方法即局部能量最大选取规则图像融合法(Image fusion is image processing of a research hot spot that is here to provide a way to select the rules of the local energy of the largest image fusion method)

说明：  ORB-SLAM 基本延续了 PTAM 的算法框架,但对框架中的大部分组件都做了改进, 归纳起来主要有 4 点: ORB-SLAM 选用了 ORB 特征, 基于 ORB 描述量的特征匹配和重定位, 都比 PTAM 具有更好的视角不变性。此外, 新增三维点的特征匹配效率更高, 因此能更及时地扩展场景。扩展场景及时与否决定了后续帧是否能稳定跟踪。ORBSLAM 加入了循环回路的检测和闭合机制, 以消除误差累积。系统采用与重定位相同的方法来检测回路(匹配回路两侧关键帧上的公共点), 通过方位图 (Pose Graph) 优化来闭合回路。PTAM 需要用户指定 2 帧来初始化系统, 2 帧间既要有足够的公共点, 又要有足够的平移量. 平移运动为这些公共点提供视差 (Parallax) , 只有足够的视差才能三角化出精确的三维位置。ORB-SLAM 通过检测视差来自动选择初始化的 2 帧。PTAM 扩展场景时也要求新加入的关键帧提供足够的视差, 导致场景往往难以扩展. ORB-SLAM 采用一种更鲁棒的关键帧和三维点的选择机制——先用宽松的判断条件尽可能及时地加入新的关键帧和三维点, 以保证后续帧的鲁棒跟踪; 再用严格的判断条件删除冗余的关键帧和不稳定的三维点，以保证 BA 的效率和精度。(Orb-slam basically continues the algorithm framework of PTAM, but most of the components in the framework have been improved, which mainly includes four points:)