基于修正Mask R-CNN算法的继电室机器人微小特征识别

Micro Feature Recognition of Relay Protection Room Robot Based on Mask R-CNN

为研究继电室机器人对继电室的实时监控效果,在继电室机器人上搭载实时监控模块并对开关柜进行智能检测。开关柜上有锁孔和开关两个微小特征需要识别,结合深度学习和双目视觉技术,对Mask R-CNN模型进行算法修正,实现继电室机器人对继电室中开关柜上微小特征的识别。

To study the effect of the real-time monitoring of relay protection room by robot,the robot was equipped with a real-time monitoring module and intelligent detection was applied to its switch gear.As the two micro features of lock hole and switch on the switch cabinet must be recognized,the Mask R-CNN model algorithm was adjusted by combining deep learning and binocular vision technology to achieve the micro feature recognition of switch gears by relay protection room robot.