文章提出了一种基于机器视觉技术的停车位状态检测方法;为了能在嵌入式系统上实现基于机器视觉的停车位状态检测,首先通过定制停车位区域,并在定制的区域内按照某种规则自动生成采样点,极大地减少了图像处理的计算资源和存储资源；其次,根据车辆驶入停车位时引起相关采样点灰度值发生变化情况进行停车位检测；最后,采用以空间高度来定制虚拟停车位,解决了相邻车位的车辆遮挡问题,通过形状匹配等算法排除路面上异常物体等各种干扰；实验结果表明,该方法不仅具有很高的检测精度和实时性,还具有较好的鲁棒性.%A parking state detection method based on machine vision is presented in this paper. In order to detect the parking state based on machine vision, firstly, computing and storage resources in embedded system are reduced in image processing through customizing the parking area and generating sampling points automatically in the custom area according to certain rules; Secondly, the parking state is detected according to the change of the gray value of relevant sampling points when vehicle come into parking) Finally, virtual parking space is customized using some space height to solve the problem of block interference between adjacent vehicles, and interferences of afferent objects in the parking are eliminated through shape matching algorithm. Experiment results show that the parking state detection algorithm has high accuracy, real time and robust feature.
展开▼
