基于Haar-like特征与Adaboost算法的前方车辆辨识技术研究  

Research of preceding vehicle identification based on Haar-like features and Adaboost algorithm

作　　者：朱志明[1] 乔洁[2] 

机构地区：[1]华东交通大学,南昌330013 [2]长安大学汽车学院,西安710064

出　　处：《电子测量技术》2017年第5期180-184,共5页Electronic Measurement Technology

基　　金：国家自然科学基金(51278062)资助项目

摘　　要：为了提高前方车辆的辨识效能,提出一种融合Haar-like特征与Adaboost算法的前方车辆辨识方法,基于海量车辆样本集进行离线训练,提取有效车辆轮廓与纹理特征,以Haar-like特征作为目标描述方法,采用Adaboost机器学习算法训练分类器,并构建特征样本级联分类器,对测试对象进行车辆存在性检测。试验结果表明,提出的融合Haar-like与Adaboost的车辆辨识算法检测准确率为91%以上,平均检测速率28ms,对车辆类型和环境干扰等非确定因素具有较强的自适应能力,提高了前方车辆纵向检测的鲁棒性,满足了车辆纵向维度的安全行驶应用需求。In order to improve the efficiency of the preceding vehicle identification, a preceding vehicle identification algorithm combined Haar-like features and Adaboost algorithm was proposed. Based on massive amounts of offline training sample set, effective vehicle contour and texture characteristics was extracted, Haar-like characteristics was used to describe the goal. Adaboost machine learning algorithms was used to trained classifier, the sample characteristics of cascade classifier was built, and the test object was used to detect the vehicle existence. The experimental result shows that the algorithm based on Haar-like features and Adaboost algorithm in the paper has a high detection rate of above 91%, and average detection speed of 28 ms, it can well adapt to the uncertain factors such as environmental interference and vehicle type, improve the robustness of the preceding vehicle detection, also meet the requirement of the safe driving in the longitudinal dimensionality.

关 键 词：HAAR-LIKE特征 ADABOOST 训练样本集 辨识 

分 类 号：TN081[电子电信—物理电子学]