CAJ | 학술논문

近年来，伴随着智能车辆的快速发展，行人检测在车辆辅助驾驶方面越来越受到重视。提出一种二维显著性纹理算子（ Two-dimensional Significant Local Binary Pattern ， TS-LBP ），该算子能够反映图像的纹理特征、显著性特征同时具有较强的抗噪声性能。首先提取感兴趣的目标区域，再提取目标区域颜色与纹理特征的融合特征对目标进行特征描述，最后采用 Adaboost 算法训练得到一个行人分类器对行人进行识别。实验结果表明，该方法检测快速准确，具有较好的检测效果。
근년래，반수착지능차량적쾌속발전，행인검측재차량보조가사방면월래월수도중시。제출일충이유현저성문리산자（ Two-dimensional Significant Local Binary Pattern ， TS-LBP ），해산자능구반영도상적문리특정、현저성특정동시구유교강적항조성성능。수선제취감흥취적목표구역，재제취목표구역안색여문리특정적융합특정대목표진행특정묘술，최후채용 Adaboost 산법훈련득도일개행인분류기대행인진행식별。실험결과표명，해방법검측쾌속준학，구유교호적검측효과。
In recent years, along with the rapid development of the intelligent vehicle, pedestrian detection in vehicle auxiliary driving is paid more and more attention. This thesis proposes a TS-LBP operator (Two-dimensional Significant Local Binary Pattern) which can reflect the texture characteristics and the significance of the image with strong anti-noise performance. The first step is to extract the interested target area. Then it extracted color and texture feature fusion of the target area for target characteristics described. Finally using Adaboost training algorithm, it got a pedestrian classifier to identify pedestrians. The research shows that the method proposed in this paper is fast and accurate with good detection results.