Design of a tomato packing system by image processing and optimization processing

In recent years, with the development of environmental control systems in plant factories, tomato production has rapidly increased in Japan. However, with the decline in the availability of agricultural labor, there is a need to automate grading, sorting and packing operations. In this research, we designed an automatic packing program with which tomato weight could be estimated by image processing and that they were able to be packed in an optimized configuration. The weight was estimated by using the pixel area properties after an L*a*b* color model conversion, noise rejection, filling holes and boundary preprocessing. The packing optimization program was designed by a 0-1 knapsack algorithm for dynamic combinatorial optimization.