摘要
文中为降低计算的复杂度,提高车型识别的效率,从整幅车辆图像中检测出车辆的车脸部分,用车脸图像对车型进行识别。采用能够快速计算的Haar-like特征,根据Haar-like特征的分布情况对其进行归一化处理,利用归一化处理后的特征构建多个弱分类器,再利用AdaBoost算法把选出的弱分类器级联为强分类器,最后用强分类器对车辆图像的车脸部分进行检测定位。实验结果表明,在100幅不同车辆图像测试样本中,车脸部分的平均检测率为79%,平均识别时间为184.98 ms。
In order to reduce the complexity of the calculation and improve the efficiency of vehicle identification,the vehicle make and model part of the whole vehicle image is detected and applied to vehicle recognition.Based on the Haar-like feature which can be calculated quickly,firstly,the Haar-like feature is normalized which is used to construct multiple weak classifiers,and then the selected weak classifiers are cascaded as a strong classifier by using the AdaBoost algorithm,and finally the strong classifier is applied to vehicle make and model detection. The experimental results show that the detection rate of the face part is 79% and the average recognition time is 184. 98 ms with 100 different vehicle image samples.
作者
朱善玮
李玉惠
ZHU Shanwei;LI Yuhui(School of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China)
出处
《电子科技》
2018年第8期66-68,81,共4页
Electronic Science and Technology
基金
国家自然科学基金(61363043)
