多传感器结合的无人机障碍物识别方法研究

随着无人机越来越多的应用在植保作业中,无人机障碍物识别正逐渐成为研究的重点.由于单一传感器识别障碍物的局限性,提出将毫米波雷达和视觉系统相结合的障碍物识别方法:利用毫米波雷达探测前方障碍物的距离,通过视觉系统检测障碍物的大小形状等信息.视觉系统结合色值提取、形态学处理等方法滤去干扰信息,得到较为光滑的障碍物轮廓,利用坐...

Study on the of method multi-sensor UAV obstacle recognition

With the increasing UAV application(Unmanned Aerial Vehicle)in plant protection operations,obstacle recognition is becoming the focus.Due to the limitation of using a single sensor for obstacle recognition,an obstacle recognition method combining the millimetre wave radar and vision system was proposed:millimetre-wave radar was used to detect the distance used to capture the size shape of obstacles.The images were processed employing colour value extraction,morphological processing and other methods to filter out the interference information.In this way,a smoother obstacle profile was obtained.Further,the transformation of coordinate system was used to realise the spatial integration of radar and visual system,the area of the obstacle is measured for the first time Synchronous thread was adopted to realise the fusion of sensors in time.According to this,it can be judged whether the UAV can pass through the obstacles smoothly.It provided the basis for obstacle avoidance decision.Experiments have proved that this method can obtain the position information of the obstacle in front with less affected by environmental factors and identify the outline and size of obstacles and calculate the distance between obstacles.