Design Of 5G Image Transmission Module For Agricultural UAV Visual Obstacle Avoidance System

With the continuous progress and development of agricultural aviation technology,Agricultural drone play an increasingly important role in agricultural production.Traditional Agricultural drone are mainly composed of support structure,batteries,rotors,flight control components,spraying components,etc.,and with the continuous development of Agricultural drone’ modular functions,image transmission modules have also become an important part of Agricultural drone.and it is required to be able to quickly and stably transmit the lens image.The binocular lens designed by the existing binocular lens manufacturer can obtain the three-dimensional information of the surrounding environment through the view difference and the geometric transformation,but it is not equipped with an image transmission module,and the image observation can only be achieved by directly connecting the computer via USB,which cannot be achieved Agricultural drone are equipped with binocular lenses for real-time image return during obstacle avoidance and feedback of obstacle information.This research aims at the characteristics of the above binocular lens,combined with 5G emerging technology,Agricultural drone load limitation and real-time requirements for image transmission screen,this study designed a 5G Agricultural drone image transmission module.The embedded development board has low power consumption,small size,high degree of integration,and good system compatibility.It can be applied to various hardware and software development environments,and it satisfies the development needs of Agricultural drone and binocular lenses.In order to realize that the real-time image of the lens can be received on the monitoring end after the binocular lens is started,this study designs a 5G network-based Agricultural drone image transmission module based on an embedded development board.The module is divided into a sending end and a monitoring end The sender includes an embedded development board,a binocular lens,and a central router.It mainly uses the V4L2 interface function of the Linux kernel and the call of the opencv function to call the binocular lens.The monitoring terminal includes a PC and a router,and mainly receives data by using a socket.The binocular lens image is collected through the program at the sending end,and the communication between the sending end and the monitoring end in the 5G network is realized under the UDP protocol,thereby achieving the transmission of the image.The video of the binocular camera is composed of continuous frame images,which can generate a depth information video stream of up to 90 fps per second,which has a large amount of information.After the binocular camera is started,the image is published in the form of a topic and cannot be directly read by Opencv.Therefore,this subject conducts in-depth research on the image transmission module of the Agricultural drone,analyzes the performance of the binocular camera currently used for the visual obstacle avoidance of the Agricultural drone,and clarifies the relevant theories involved in the design process of the image transmission module.Demonstrate the design scheme of the image transmission module,including the choice of the development board,the choice of the embedded operating system,the construction of the software platform,the use of the binocular camera,the selection of the transmission scheme and the overall realization of the image transmission module.According to the design requirements,two different transmission methods are designed,one is the direct transmission of video frames using the C++programming graphical interface on the QT creator development platform,and the other is the video frames according to the resolution in the C++ environment.Split and then transfer.Finally,a secondary development of the binocular camera startup program under the ROS system was carried out.The image transmission of the binocular camera under the 5G network was realized by the fusion of the two transmission methods,which provided technical support for the visual obstacle avoidance of Agricultural drone.Through the experimental verification of the two transmission methods under the 5G network,the experimental results show that: the transmission method of image segmentation can control the image transmission delay of the binocular camera within25 ms.It satisfies the monitoring of the Agricultural drone screen,and realizes that after the binocular camera is started,the monitoring end can receive the image of the binocular camera.It meets the design requirements of the image transmission module and has good real-time performance and stability.