Research On Medium-sized Apple Post-harvest Treatment Technology

Since the 14 th Five-Year Plan,the state has strongly supported the development of agricultural science and technology,and as an important part of agricultural modernization,the apple processing industry has received continuous attention from the state in recent years.Although China’s apple planting area and production are among the top in the world,there is a gap between China’s apple post-harvest processing equipment and that of foreign countries,resulting in a large amount of labor costs and a lower commercialization of apples.In order to improve the market competitiveness of our apples,we need to continuously invest more research in apple post-harvest processing equipment.At present,the degree of automation of apple post-harvest processing in China is low,and the man-made mechanical damage is serious;the grading part has low precision and poor detection of small defects;the packaging based on heat-shrinkable film packaging cannot protect apples in transit,which seriously restricts the industrialization of apples.This paper addresses this problem of apple post-harvest processing,and designs an intelligent device for medium-sized apple post-harvest processing based on traditional equipment,using theoretical analysis,design calculation,simulation verification,and YOLOv5 algorithm improvement.(1)Analyze the current research status of traditional apple post-harvest cleaning,grading and packaging equipment,summarize the shortcomings and improvement goals of traditional equipment,and determine the main research content as three major modules: loading and cleaning equipment,quality grading equipment,and packaging equipment.(2)According to the improvement target,the structural design of each component of feeding,lifting,cleaning,wiping and drying as well as fruit collection and feeding is carried out,and the model of automatic feeding and cleaning equipment is established by using industrial 3D software to design a fruit box automatic turning and feeding device to realize automatic apple feeding and fruit box collection;and the parameters of each power driving device,nozzle structure type and installation position as well as the installation height of wiping device are determined by the analysis and calculation of each component.The installation height of the wiping device is determined by analyzing and calculating each component,and finally the feasibility is verified by using engineering simulation software.(3)To solve the problem of incomplete detection information of the grading system,we design a kind of fruit-bearing cup based on TRIZ theory,which can achieve full-angle rotation of apples during the working process,and design the fruit-unloading device for the fruit-bearing cup to realize automatic grading and fruit unloading;to ensure that the surface of apples is not damaged during fruit unloading,we design experiments to determine the surface material and installation height of the receiving plate,and use the engineering simulation software to evaluate the surface of the fruit-bearing cup.In order to ensure the damage of the surface of the fruit unloading,we designed the experiment to determine the surface material of the receiving tray and the installation height,and used the engineering simulation software to verify the fruit-bearing capacity,the stability of conveying and the fruit unloading process of the accompanying conveying cup.(4)To design a foam-based net sleeve packaging machine based on the physical and chemical characteristics of apples,which can realize the continuous packaging of single fruit and reduce the collision damage of apples during transportation through the mutual cooperation of hook and net sleeve support mechanism.(5)For the current situation of apple fruit diameter and defect detection and grading,by analyzing the grayscale distribution of apples under different channels of RGB and HSI,the H channel with more obvious contrast is selected for further processing of images,and the minimum outer circle method is selected for fruit diameter measurement,and the diameter measurement error is within 5%,which meets the grading requirements.(6)To address the problems of slow detection speed and poor detection ability of small target defects by traditional YOLOv5,the benchmark network framework was improved and ASFF and CBAM network mechanisms were introduced.The final model test result of m AP was 0.88,P was 0.856 and FPS was 35.8,which met the requirements of online detection.(7)Design the whole control system of apple post-harvesting processing equipment,determine the controller parameters,and build a simulated experimental bench for debugging tests.