基于OpenMV图像识别的自主迷宫寻宝小车

Autonomous maze treasure hunting car based on OpenMV image recognition

基于机器视觉模块OpenMV和TC264单片机,设计基于图像识别的自主迷宫寻宝小车。通过识图装置OpenMV识别藏宝图,定位宝藏坐标、识别迷宫矩阵,并通过串口发送给单片机,利用多目标A算法进行路径规划,利用陀螺仪与编码器进行实时定位,根据小车当前的坐标与宝藏坐标对比来进行小车的控制,同时使用超声波避障模块及按键等实现迷宫寻宝的任务。通过实验证明,小车系统运行流畅,各个模块之间协调配合,定位精度可达0.2 mm,判别真伪宝藏并将信息发送给单片机的时间在1~2 s,在科学性、准确性、可行性、创新性和完整性方面均有较好的表现。

On the basis of the machine vision module OpenMV and microcontroller unit(MCU)TC264,an autonomous maze treasure hunting car is designed based on image recognition.The map reading device OpenMV is used to identify the treasure map,locate the treasure corrdinates and identify the maze matrix.After the identification,the messanges of the maze and treasures are sent to the MCU by the serial port.The path is planned with the multi⁃objective A algorithm.The gyroscope and the encoder are used for real⁃time positioning.The car is controlled according to the comparison between the current coordinates of the car and the treasure coordinates.Meanwhile,the ultrasonic obstacle avoidance module and keys are used to achieve the tasks of finding treasures in the maze.Experiments show that the small car system operates smoothly,with coordination and cooperation among various modules.Its positioning accuracy is 0.2 mm.It takes about 1 s to 2 s for the hunting car to judge the authenticity of the treasure and send the information to the MCU.The results show that the car has good performance in terms of scientificity,accuracy,feasibility,innovation and completeness.