Research On Intelligent Control System Of Vegetable Seed Coating Machine Based On Computer Vision

Seed coating as a key part of seed processing,act as an important role in terms of improving the resistance to pests and diseases and promoting seed germination.The results of seed coating treatment also have a great impact on seed quality.Since the promotion of seed coating technology in China,seed coating technology has been rapid developed.The assessment of coating results for traditional coating machine relies largely on professional operators.And the adjustment of operating parameters depends entirely on the operator’s experience.In order to ensure and improve the quality of coated seed and lower the requirement of coating operators,intelligent seed coating machine has to be designed.With rapid development in recent years,computer vision technology has been widely used in agriculture applications.In this paper,considering the characteristics of coated seeds and the technical requirements of coating treatment,the method of computer vision was used to identify the results of seed coating,and further research on relationship between the coating parameters and seed coating results was conducted,which provides the theoretical basis and technical support for the on-line automatic adjustment of coating parameters.Then intelligent detection and control system of vegetable seed coating machine based on computer vision is put forward for the first time.In the study of coating result identification algorithm,the preprocessing and segmentation methods of seed coating images are studied.In order to solve the problem of adhesion in seed coating images,the watershed segmentation algorithm was used in the local area.The film encapsulation rate,seed color and texture characteristics was extracted as features of coated seed.And then support vector machine classifier was used to classify coated seed.Based on the classification of the coated seed and the requirement of working parameters,this paper proposed a method to classify the coating results into multiple classes based on the support vector machines and decision trees.According to the classification results,the method of working parameter adjustment can be proposed.Based on the above points,an intelligent control system of vegetable seed coating machine based on machine vision was constructed.At last,experiments were conducted with pepper seed and BY-150 coating machine to verify the feasibility and adaptability of the algorithm.