【摘要】 采用单一传感器识别前方车辆易造成误判,为了提高检测准确率,提出了一种基于激光雷达信息和单目视觉信息的车辆识别方法,该方法综合考虑了激光雷达传感器和单目视觉传感器的信息。首先对激光雷达原始信号进行预处理,然后对筛选后的数据进行最临近距离法聚类处理,并转换到图像坐标系中,初步确定障碍物检测的ROI (感兴趣区域)。提取ROI区域的图像,并对其进行灰度化,灰度增强和图像滤波的预处理。通过计算预处理后图像的熵值归一化对称性测度,完成前方车辆的检测。通过实验,验证该方法可以较好地识别前方车辆,弥补了单一传感器在车辆识别中的不足,同时耗时较短,可以满足实时性的要求。更多还原

【Abstract】 Identification of vehicles ahead with a single sensor is easy to cause error detection. In order to improve the accuracy of the detection, we propose a vehicle detection method based on laser radar and monocular vision. The method takes into account the information of the laser radar sensor and the monocular vision sensor. The radar raw signal is preprocessed firstly, and then the filtered data is processed by clustering and converted to the image coordinate so as to preliminarily determine the region of interest(ROI) of vehicle detection. The image of ROI is extracted and preprocessed by graying, gray enhancement and image filtering. And then a symmetry measurement of normalized entropy of the preprocessed image is obtained to detect preceding vehicles. It is verified that this method can well identify vehicles ahead,which compensates for the shortage of single sensor by experiment; and the time consuming is short, which can meet the real-time requirement.更多还原