Research On Obstacle Avoidance Technology Of Plant Protection UAV Based On Structured Light Vision

In recent years,the industry of China's plant protection UAV has rapid development.The way of using UAV for plant protection is more and more common.China's farmland operating environment universally exists telegraph poles,trees and other obstacles.When the flying UAV encounters obstacles,it must rely on operator's visual observation and manual remote control to avoid obstacles.This way of artificial obstacle-avoidance seriously affects the operation efficiency,and threatens UAV flight safety.Therefore,the study of automatic obstacle avoidance technology is of great significance for the safe flight and fully automatic flight of plant protection UAV.The main work of this paper is:(1)For the obstacle avoidance of plant protection UAV,an obstacle detection method based on structured light vision was proposed.Based on the principle of triangulation,through the special optical path design between the semiconductor laser and CCD sensor,an optical detection system to detect front obstacle information was designed.The optical system consisted of line structured light source,CCD,lens and filter.The angle between the optical axis of the light source and the CCD imaging optical axis was designed as ‘?'.The distance between the light exit hole of light source and the CCD imaging optical axis was designed as ‘z'.The line structured light emitted by the laser was reflected by the surface of the obstacle.It was imaged on the CCD target surface.Through the image acquisition,processing and calculation,the information of distance,azimuth and width of front obstacle could be extracted.Finally,combining with the attitude parameters of the UAV flight,the method to correct the test results according to the pose angle of the fuselage was proposed.(2)Hardware and software design of the obstacle avoidance system were completed.Choosing multi-rotor wing plant protection UAV as the obstacle avoidance test platform,small size IPC as the obstacle avoidance control module.According to the obstacle avoidance requirements of the plant protection operation and the view field of the fuselage size,an obstacle avoidance detection module was designed,composing of a CCD,a lens,a structured light source,a filter and a module bracket.Thus the design of the hardware platform was completed.The software design was divided into three parts: obstacle avoidance pre-flight preparation,obstacle avoidance detection and obstacle avoidance action.The part of the pre-flight preparation maintained obstacle avoidance parameter settings and the before-fly self-test.The detection part included the design of real-time detection process and image processing.The image processing method of classifying obstacles was proposed.For the motion module,according to the relative position of the obstacle and the UAV,the corresponding motion command and movement distance were generated and sent to the flight controller,then flight points of obstacle avoidance were inserted by the flight controller.Finally completing the software design to avoid obstacles.(3)Calibration design for the parameters ‘?' and ‘z' of the detection module was completed.According to the principle of distance detection,a calibration method of calculating the ‘?' and ‘z' to be calibrated by the setting distance was designed,and a suitable calibration plate was designed.Finally,through the hardware and software design of the calibration system,the calibration for parameters of the detection module was completed,ensuring the detection accuracy of the detection module.(4)For this obstacle avoidance system,the calibration experiment,the static and dynamic experiment of obstacle detection and the obstacle avoidance flight experiment have designed.Experimental results showed:within 3000 mm,the detection module of the distance detection deviation does not exceed ±60mm,azimuth detection deviation does not exceed ±4.0°,width detection deviation does not exceed ±10mm.The automatic obstacle avoidance during flight was realized.The effectiveness of the obstacle avoidance technology for plant protection UAV based on structured light vision is verified.