使用opencv_traincascade训练Haar、HOG、LBP Adaboost分类器

opencv_traincascade.exe默认表格

使用opencv_traincascade训练Haar、HOG、LBP Adaboost分类器_第1张图片


训练正样本2400,负样本3300

1.opencv_traincascade.exe  -data traincascade -vec pos.vec -bg neg.txt -numPos 2000 -numNeg 3300 -numStages 20 -precalcValBufSize 200 -precalcIdxBufSize 1000

 -featureType LBP -w 18 -h 24 -bt GAB -minHitRate 0.95 -maxFalseAlarmRate 0.5 -weightTrimRate 0.95 -maxDepth 1 -maxWeakCount 100

如果出现 Parameters can not be written, because file traincascade/params.xml can not be opened 错误

则自己需要手动创建一个文件夹 traincascade

2.如果出现如下问题.

Traincascade Error:Bad argument(Can not get new positive sample.The most possible reason is insufficient count of samples in given vec-file.

记得一定要numPos小于vec_file文件里面的数 一般numPos为0.9*num_in_vec或者为0.8*num_in_vec

具体讲解可以参考 http://answers.opencv.org/question/4368/traincascade-error-bad-argument-can-not-get-new/


你可能感兴趣的:(OpenCV)