▍1. Calculate_F_matrix
立体视觉 输入匹配点坐标,计算基础矩阵,含多种方法(Computer Vision, Calculate the foundation matrix by various methods )
立体视觉 输入匹配点坐标,计算基础矩阵,含多种方法(Computer Vision, Calculate the foundation matrix by various methods )
使用Opencv的videocapture功能,处理视频,添加水印,创建了原视频的9画面切割显示,QT使用了相应的布局和多线程,对于初学者有很好的帮助。(Using the videocapture function of Opencv, processing video, adding watermark, creating the 9 picture cutting display of the original video, QT uses the corresponding layout and multi thread, which is very helpful for the beginner.)
说明: 基于SVM的烟雾识别系统 运动区域提取使用的是 自适应混合高斯背景建模 特征算法使用的是 HOG+LBP 机器学习方法使用的是 scikit的SVM(Smoke recognition system based on SVM The moving region extraction uses adaptive Gaussian mixture background modeling feature algorithm, hog + LBP machine learning method and scikit SVM)
人脸及行人检测中在opencv中要用到的haarcascade_righteye_2splits.xml,即特征检测的haar分类器,包括正面,侧边,上下左右各种,很全面(Face and pedestrian detection in opencv use to haarcascade_righteye_2splits.xml, feature detection haar classifier, including front, side, up and down various, very comprehensive)
说明: 图像修复 实现以c++和opencv为环境,修补图像的算法(image inpainting . based on c++ and opencv to restore the picture.)
旋转向量与旋转矩阵转化(罗德里格斯 ) 根据OPENCV里的 cvRodrigues2()函数原理编写的纯c的函数,不需要依赖OPENCV(Rotation vector and rotation matrix transformation (Rodriguez))
说明: 根据左右相机拍摄的图片,先进行单目标定,再进行双目标定,之后通过立体校正和立体匹配得到视差图,计算深度(According to the pictures taken by the left and right cameras, monocular calibration is performed first, and then binocular calibration is carried out. Then the parallax map is obtained by stereo correction and stereo matching, and the depth is calculated)
说明: 通过代码实现对图片和视频中车辆的检测与识别(Through the code to achieve the detection and recognition of vehicles in pictures and videos)
说明: 标准粒子滤波算法,用于地形匹配,对初学者具有参考价值(Standard particle filter algorithm for terrain matching)
说明: 基于opencv的激光线中心提取源码,需要配置下自己的opencv和图像文件路径,可以直接运行,输出激光线的中心(The source code of laser line center extraction based on OpenCV needs to be configured with its own opencv and image file path, which can run directly and output the center of laser line)
说明: 采用传统图像处理方法,能够实现对车道线的检测,检测效果较好(The traditional image processing method can realize the detection of lane line, and the detection effect is good)
说明: 使用opencv可以读取bmp肺部CT图片并准确分割出肺部的轮廓。(we can read BMP lung CT images and accurately segment the lung contour by opencv.)
基于VC++ 和opencv的焊盘图像 内径外径的测量。包括滤波、边缘检测、轮廓提取和面积计算等算法。包含原图片(Based on the measurement of the pad image VC++ opencv Inside Outside. Including filtering, edge detection, contour extraction and area calculation algorithm. Contains the original image)
说明: 此代码简单易学,能有效检测烟雾,识别率高,机器学习。(This code is easy to learn, can effectively detect smoke, high recognition rate, machine learning.)
车道线检测,在vs2010配置opencv的情况下亲测可用,代码没有做出透视变换(Lane detection, in the case of pro-test opencv vs2010 configuration is available, the code does not make a perspective transformation)
matlab程序进行图像分割,图像的多阈值分割(Matlab program for image segmentation, image multi-threshold segmentation)
说明: matlab程序进行图像分割,图像的多阈值分割(Matlab program for image segmentation, image multi-threshold segmentation)
开发环境:vc6.0 ,opencv1.0.整体功能:实现车牌跟踪,检测,车牌定位,车牌分割,字符 分割,字符识别,其中车牌跟踪和车牌检测不是很好,但是对于已经拍好的图片,车牌定位, 车牌分割,字符分割和字符识别效果很好。程序中的核心算法,本人已加详细注释。具体详见程序说明。希望对你有所帮助。(Development Environment: vc6.0, opencv1.0. Overall functions: license plate tracking, detection, license plate location, license plate segmentation, character segmentation, character recognition, including the license plate license plate tracking and detection is not very good, but has take a good picture, license plate location, license plate segmentation, character segmentation and character recognition with good results. The core algorithm in the program, I have added detailed comments. See detailed description of the procedures. I hope for your help.)