用于违章停车车辆自动取证的四轴飞行器系统

A quadcopter system for illegally parked vehicle inspection

摘要:

借助于计算机视觉技术实现了一种基于四轴飞行器的违章车辆取证系统。四轴飞行器在巡检路段上空飞行过程中，通过训练好的支持向量机(SVM)车辆检测算法自动检测视野中的违章停车车辆。如违章停车车辆存在，四轴飞行器将根据视觉处理结果自动调整姿态，便于摄像头以合适的尺度和角度识别车牌；最后通过模式识别技术捕捉视野中车辆的车牌并向地面站传输证据。与传统的人工巡检和基于道路固定摄像头的方案相比，本系统更灵活、高效，可以作为交通智能化管理新的解决方案。

Abstract:

A quadcopter system for illegally parked vehicle inspection is implemented by means of computer vision technology. When the quadcopter patrols a specific area，the system automatically detects illegally parked vehicles in vision through the pre-trained Support Vector Machine(SVM) detector. If the vehicle exists，the quadcopter adjusts its own attitudes to an appropriate place according to the processing results in order to recognize the plate license. The pattern recognition technology is applied to extract the plate license and transmit it to the ground station. Compared with the traditional manual inspection method and algorithms based on fixed cameras，the proposed system is more flexible and efficient，which can serve as a new solution to the intelligent traffic management.