Design of Greenhouse Tomato Picking Robot Based on intelligent control

At present, China's Tomato Picking methods mainly rely on labor, which has problems such as high labor intensity, high labor cost and high fruit damage rate. This paper designs a small greenhouse tomato picking machine, including visual recognition system, lifting system, picking system, storage system, automatic driving system, power supply system, control system and other parts. It is convenient for operation, use and maintenance, in order to realize the mechanization and automation of tomato picking process, reduce the picking cost and improve the work efficiency. In the greenhouse dynamic environment, the control variable method is used to control the light environment variables. A total of 8 rounds of tests were conducted, and each round of tests lasted for 5 minutes. Each round of test was conducted in the ridge section with different tomato distribution density to evaluate the universality of the system. The test scenarios include branch and leaf occlusion and artificial simulation of dynamic obstacles. The final result is taken as the average value of 8 rounds of test results. The core positioning scheme uses LIDAR and binocular vision. Through multi-level data fusion and dynamic interference suppression technology, the scheme improves the positioning accuracy of the system in the signal interference environment, and improves the perception ability in complex environment. Combined with the visual assisted LIDAR positioning scheme, the picking efficiency is improved by 15%. The false cutting rate is stable at 6.5%, and the developed picking prototype has reached an average picking rate of 28.4 pieces/minute, which basically meets the needs of automatic tomato picking. It has the advantages of low cost, integration of harvesting and collection, adaptability to different ridges, continuous operation throughout the day and so on. It can effectively assist fruit farmers, reduce labor burden and cost, and has a good application prospect.