| The intelligent image acquisition platform can realize real-time image acquisition.However,due to factors such as dynamic changes in the growth of greenhouse crops and relatively narrow roads,traditional manual or fixed camera acquisition methods have problems such as high labor intensity,single shooting angle,and small amount of information obtained.If a mobile platform is used for image acquisition,the magnetic navigation layout and maintenance costs in the navigation implementation method are high,and there are many crops in the greenhouse and the light is not good,and visual navigation is not easy to realize.Therefore,this research intends to design an image acquisition mobile platform that can realize autonomous navigation,and improve the positioning effect of the greenhouse mobile platform through the fusion positioning method of UWB(Ultra-Wideband,ultra-wideband)and LIDAR(Light Detection and Ranging,laser radar).The main work of this paper is as follows:(1)The overall design of the image acquisition mobile platform: Based on the greenhouse environment,the main structure of the image acquisition mobile platform was built.The lifting design can better adapt to different plant heights,and then complete the hardware selection design and control system design for subsequent development A solid foundation has been laid.(2)Mobile platform positioning: UWB and IMU(Inertial Measurement Unit,inertial measurement unit)fusion positioning method is adopted to reduce the non-lineof-sight error caused by UWB positioning due to crop occlusion,and improve positioning accuracy;secondly,an odometer is established The motion model and LIDAR observation model discuss the AMCL(Adaptive Monte Carlo Localization)algorithm,laying the foundation for fusion positioning.Finally,based on the AMCL algorithm,two weight updates are used to realize UWB and LIDAR fusion positioning,which can effectively improve its positioning accuracy.(3)Navigation of the mobile platform: Based on the Cartographer algorithm,the laser SLAM environment mapping is completed;combined with the requirements,the path planning is divided into global path planning and local path planning.In the global path planning,A^* algorithm and Dijsktra algorithm are compared and analyzed,the A^*(A-star)algorithm with higher search efficiency was selected;in the local path planning,the DWA(Dynamic Window Approach)algorithm was adopted in consideration of the real-time performance and the memory size of the algorithm.(4)Experiment and analysis of the greenhouse environment: based on the built image acquisition mobile platform,the creation of a greenhouse map based on lidar is realized.The created map has a high degree of restoration and clear boundaries.In the UWB positioning optimization experiment,the average offset error of pure UWB positioning is 0.32 m,while the error of fusion positioning is 0.07 m,and the effect of fusion positioning is improved by 25% compared with that,reaching the expected goal.In the experiment of fusion positioning and navigation,the mobile platform can plan the path according to the target point and conduct experimental navigation.The driving offset error of each section of the path is less than 15 cm,which meets the requirements of greenhouse navigation.During the navigation process of the mobile platform,the acquisition frequency of the camera module is set to 6,and the quality of the pictures collected in the two road environments is evaluated,from the average gray level,image entropy and perceptual sharpness index(PSI)three The average grayscale is about 45%,the image entropy is about 7.5,and the PSI fluctuates around 0.5.The results show that the image brightness is medium,contains more information and details,and the outline is clear and the image quality is high.It can be used for Subsequent image processing and other work. |
PDF Full Text Request