卢胜男, 李小和. 基于对称FAST特征的车辆目标检测方法[J]. 微电子学与计算机, 2020, 37(2): 37-42.
引用本文: 卢胜男, 李小和. 基于对称FAST特征的车辆目标检测方法[J]. 微电子学与计算机, 2020, 37(2): 37-42.
LU Sheng-nan, LI Xiao-he. Vehicle detection method using symmetrical FAST feature[J]. Microelectronics & Computer, 2020, 37(2): 37-42.
Citation: LU Sheng-nan, LI Xiao-he. Vehicle detection method using symmetrical FAST feature[J]. Microelectronics & Computer, 2020, 37(2): 37-42.

基于对称FAST特征的车辆目标检测方法

Vehicle detection method using symmetrical FAST feature

  • 摘要: 针对动态复杂环境下车辆目标识别和定位问题,提出一种基于对称FAST特征的车辆目标识别方法.利用车辆的对称特征,采用水平镜像矩阵构造FAST特征描述矢量,检测具有对称特性的FAST特征点对,以此为车辆特征聚类的约束条件,确定车辆中心线位置,并结合车辆底部阴影特征对车辆目标进行识别和定位.实验结果表明,该方法提取的目标车辆区域位置准确,平均检测准确率高达93.2%.

     

    Abstract: Aiming at the problem of moving vehicle detection and recoginiton, this paper proposed a vehicle detection algorithm based on vehicle symmetrical FAST feature. Based on the vehicle symmetrical feature, a novel symmetrical FAST points extraction algorithm is presented by constructing FAST descriptor. Then the vehicle symmetrical features are extracted using this method, which is also used to locate the central position of vehicle. At last, the vehicle will be recognized combining with the shadow features of the target. The experiment results show that the proposed method can detect vehicle area accurately, which can reach an average accurate detection rate of 93.2%.

     

/

返回文章
返回