摘要
提出一种基于单眼视觉和超声波测距的树莓派智能机器人车检测静态和动态障碍物的方法.采用改进的单眼视觉障碍物检测算法,对室内的静态和动态障碍物进行轮廓检测,并利用超声波传感器测量机器人车与障碍物之间的距离.针对静态障碍物检测,在图像预处理阶段引入图像增强,并通过HSV图像提取不同障碍物颜色特征,以提高障碍物轮廓标定的效率和准确率.针对动态障碍物检测,结合背景差分与3D图像显示技术实现动态目标捕捉,并设置距离决策模块记录障碍物位置信息.试验结果表明,该方法可有效减少障碍物检测的平均消耗时间以及障碍物位置信息的错误率,提高室内障碍物检测的效率和准确性.
A method based on monocular vision and ultrasonic ranging for the intelligent Raspberry Pi robot for detecting static and dynamic obstacles was proposed.An improved monocular visual obstacle detection algorithm was applied to perform contour detection on indoor static and dynamic obstacles,the distance was measured between the robot car and obstacles with an ultrasonic sensor.For static obstacle detection,image enhancement was introduced in the image preprocessing stage,and different obstacle color features were extracted through HSV images to improve the efficiency and accuracy of obstacle contour calibration.For dynamic obstacle detection,background difference was combined with 3D image display technology to achieve dynamic target capture,and a distance decision module was set up to record obstacle location information.The experimental results show that the method can effectively reduce the average consumption time of obstacle detection,and improve the accuracy of indoor obstacle detection.
作者
唐明
李媛媛
TANG Ming;LI Yuanyuan(School of Electronic and Electrical Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)
出处
《上海工程技术大学学报》
CAS
2022年第1期69-76,共8页
Journal of Shanghai University of Engineering Science
关键词
单眼视觉
超声波传感器
室内障碍物检测
图像增强
背景差分
monocular vision
ultrasonic sensor
indoor obstacle detection
image enhancement
background differencing
