OpenCV下嵌入式轨道检测小车图像采集处理系统研究与设计

Research and Design of Image Acquisition and Processing System for Embedded Track Detection Vehicle Based on OpenCV

在线路基础设施建设中,轨道安全维护是急待解决的薄弱环节,而轨道状态检测是保障列车安全运行的重要手段;应用OpenCV图像处理集,搭载嵌入式计算平台,对轨道图像采集处理,快速的分割并识别出轨道图像中的关键设备状态,是解决现有轨道检测技术成本高,效率低的一个重要研究课题;为研究该课题,设计了嵌入式轨道检测小车图像采集处理系统,包括整体系统结构设计、硬件选型与电路设计、软件运行结果测试;选用Linux平台作为处理核心,用Cadence设计STM32单片机同步采集触发电路,设计供电系统电路、优化正交编码电路与信号调理电路,在μC/OS-Ⅱ系统下编写同步触发采集程序,完成小车采集系统配置安装与测试,验证了OpenCV程序在嵌入式平台上的工作稳定性以及算法的实用性。

In the construction of railway infrastructure, track safety maintenance is a weak link to be solved urgently, and track condition detection is an important means to ensure the safe operation of trains. The application of OpenCV image processing set and embedded computing platform to track image acquisition and processing, fast segmentation and recognition of key equipment status in track image is an important research topic to solve the high cost and low efficiency of existing track detection technology. In order to study this subject, an embedded image acquisition and processing system for track detection trolley is designed, including the overall system structure design, hardware selection and circuit design, and software operation results test. Using Linux platform as the processing core, using Cadence to design STM32 single chip synchronous acquisition trigger circuit, design power supply system circuit, optimize orthogonal coding circuit and signal conditioning circuit, write synchronous trigger acquisition program under the system of C/OS-II, complete the configuration, installation and testing of car acquisition system, and verify that OpenCV program is embedded. The stability of the platform and the practicability of the algorithm are verified.