Research On The Key Technology Of The Image Navigation In The Intelligent Inspection Of The UAV

The traditional inspection of water transmission pipeline mostly adopts manual inspection,which has the following two shortcomings: first,the inspection workload is large,the inspection time is long,resulting in waste of resources.Most of the second water pipelines are located in remote areas,and the environment is relatively poor,so there may be potential safety hazards for inspectors.Therefore,using UAV as the inspection tool of water transmission pipeline can not only effectively reduce the workload of inspection personnel,save the cost of pipeline inspection,but also avoid the damage to inspection personnel caused by the harsh environment in the field.In order to realize the UAV pipeline inspection,the most important thing is to find a reliable,safe and low-cost UAV navigation mode.At present,the commonly used navigation systems of UAV include: GPS,Doppler navigation system,inertial navigation system,image navigation system,and the navigation mode of combining multiple navigation systems.However,most of the water conservancy facilities are located in remote areas,where GPS positioning accuracy is low and there is no landmark building,so the traditional navigation method will not only increase the cost of inspection,but also increase the risk of UAV out of control caused by low navigation accuracy and other reasons.In this paper,the idea of UAV inspection of water pipeline based on new image navigation is proposed.The image navigation algorithm proposed in this paper is to determine the pipeline angle offset and position offset by image filtering,different space threshold segmentation,ROI(region of interest extraction)method,perspective change and other related methods according to the difference of the target and background color,so as to achieve the attitude adjustment of UAV.The patrol inspection method mainly includes the following three aspects: first,it no longer realizes the location through the matching results of the real scene and the map or landmark stored in memory in advance,so it effectively saves memory space.The second is to use the bramble Raspberry Pi 3B as the image processing platform,which has the advantages of high processing speed,large memory,small size,easy to carry,integrated with high-definition camera and so on.It avoids the waste of resources,large volume,poor mobility and other problems caused by the PC as the image processing platform.In the third part,ATK-F405 is used as an open source flight control system,and ATK-F405 is developed for the second time.Finally,the intelligent inspectionfunction of water pipeline based on UAV is realized.After experimental research,the image navigation algorithm proposed in this paper can realize the calculation of pipeline angle offset and position offset,and initially realize the flight attitude control of UAV in ATK-F405 flight control.The research results of this paper lay the foundation of image navigation and flight control algorithm for the later realization of intelligent patrol of water pipeline based on UAV.