基于多视角机器视觉的球形水果无损检测与分选系统研究OA

To overcome the drawbacks of traditional fruit sorting methods,such as low efficiency,heavy reliance on manual labor,and high costs,citrus fruits were selected as the research object,and a non-destructive detection and automated sorting system based on multi-view machine vision was developed,aiming to promote the intellectualization and standardization of post-harvest fruit processing.Equipped with multiple CMOS industrial cameras evenly arranged along the circumferential direction,multiple images were synchronously captured during the rolling process of the fruits by the system,which fully covered the fruit surfaces and effectively addressed the problem of defect omission caused by single-view imaging.By fusing multi-view image information,a comprehensive multi-index evaluation system was established,which included sphericity,weight(with a prediction error≤10%),maturity and surface defects,thus enabling the accurate detection of both the external appearance and internal quality of the fruits.Based on the aforementioned features,a multi-index fusion grading mathematical model conforming to the national grading standards was constructed,and an automated device integrating the functions of feeding,imaging,decision-making and sorting was developed.Experimental results showed that the sorting speed of the system reached 3～5 fruits per second,and the recognition accuracy for obvious surface defects exceeded 90%.The system provided an efficient and reliable automated solution for the post-harvest quality detection and grading of spherical fruits,and offered practical technical support for improving the quality and efficiency of the fruit industry as well as realizing its full-process mechanization.