搜索资源列表
Vehicle1.3
- VC6.0 + QT4 + OpenCV 环境。演示视频地址http://v.youku.com/v_show/id_XMzA3MTU2Mzky.html (低角度车流量检测器)一种基于车头特征的车辆检测，很好的应用于低角度摄像机。与地感线圈车检器相比，解决了跟车问题。与传统基于背景建模的检测方法相比，解决了道路拥堵问题并解决了传统方法受限于摄像机必须架设很高的问题。同时能有效去除自行车、摩托车、行人的干扰，准确率可达95 。 all rights reserved by birdwcp-V
Traffic-statistics
- 能实现运动检测与跟踪，不统计出路面的车流量，准确率在90 以上-Traffic statistics
opencv-sanjiao
- opencv 从背景图像中找三角形，矩形，注意要用Png格式准确率就100 -opencv from the background image to find triangles, rectangles, and pay attention to use Png format accuracy of 100
yanse
- 基于c++开发，能够调用电脑的摄像头拍照或者调用现存照片。然后进行固定颜色的提取，准确率高，适合初学者。-Based on c++ development, be able to call the computer' s camera to take photos or call an existing photo. Then fixed color extraction, high accuracy, suitable for beginners.
car
- 在工业图像下的一种车辆定位系统,具有较高的准确率-In industrial image under a vehicle positioning system with high accuracy
faceLBP
- 包含一个性别分类的cpp文件，和已经训练好的gender.yml模型，最后计算出样本分类准确率。建议使用AR人脸数据库。使用格式为genderLBP 训练样本.txt 测试样本.txt。因为不用训练模型了，测试样本随便写个行了。测试样本的格式是：路径 1或者0。1表示男性，0表示女性。-Cpp file contains a gender classification, and has trained gender.yml model, and finally calculate the sam
HOG_OpenCV
- HOG即histogram of oriented gradient, 是用于目标检测的特征描述子，该技术将图像局部出现的方向梯度次数进行计数，该方法和边缘方向直方图、scale-invariant feature transform类似，不同的是hog的计算基于一致空间的密度矩阵来提高准确率。Navneet Dalal and Bill Triggs首先在05年的CVPR中提出HOG，用于静态图像or视频的行人检测。-HOG i.e. histogram of oriented gradien
PORKER
- 一个利用模板匹配实现的扑克牌花色和大小识别程序。开发工具是opencv2.4.2+vs2010,准确率较高。-One using the template matching color and size to achieve poker identification procedures. Development tools are opencv2.4.2+vs2010, higher accuracy.
plate
- 基于opencv的车牌识别源码，准确率应该不错，用C语言写的。-Opencv source code based on license plate recognition accuracy rate should be good, written in C language.
kalman2
- 卡拉曼跟踪实例,跟踪准确率较高，效果较好，只是单纯地跟踪，不能输入图像信息-kalman tracking example
facedetection2
- 用opencv库，在vs2013上面跑成功的人脸检测工程包，亲测，准确率还是阔以的-Using opencv library in vs2013 running above successful face detection works package, pro-test, the accuracy rate is still wide in the
cardetect
- opencv运动模板检测车辆，可以检测出运行方向，准确率较高-Opencv motion template detection vehicle, can detect the running direction, the accuracy rate is higher
Vehicle-Identification
- 多目标移动车牌能准确识别，比之前车牌准确率更高-Multi-target mobile license plate can accurately identify a higher accuracy rate than before the license plate
手势识别
- 收拾识别，简单实现，识别准确率不是很高，入门学习(Clear identification, simple implementation)
TestLBP
- 使用vs2008开发平台，c++语言，对人脸图片进行LBP特征提取，再通过距离度量计算人脸相似度，在100人的小库中准确率超过80%(We use VS2008 development platform and c++ language to extract LBP features from face images, and then calculate face similarity by distance metric. The accuracy is over 80% in 100 sm
zbar-master
- 基于java编写的zbar二维码识别库。亲测可用。识别速度快，准确率较高(ZBar QR code recognition library based on Java. Pro test available. High recognition speed and accuracy)