Research Of Apple Size Classification Technology Based On Machine Vision

With the improvement of living standards of society, people’s health awareness of self is also growing; the fruit is rich in many nutrients and easy to absorb, so the world’s increasing demand for fruit year after year, and on the fruit quality requirements more and more stringent. Our fruit acreage has expanded each year, the production also will increase the amount of fruit, but its sale price is generally relatively low, can not lead to higher income. The current situation makes the fruit sorting work has become very important, especially for the size grading work. In this paper, machine vision technology, the fruit size fractionation techniques are studied. The main contents are:(1) Analysis of the current situation and development trend of domestic and foreign fruit size intelligent detection technology, studied the working principle of the machine vision system is proposed for the actual situation of the unit machine vision systems solutions:image acquisition system, determine the various hardware models and build a hardware platform machine vision systems; data analysis systems, selected as a machine vision software HALCON software development platform, the preparation of the associated application, developed an electromechanical device for size grading work..(2) The relative sizes of apple grading techniques. As referred to herein refers to the relative size of the pixel size, which is based on the number of pixels of Apple picture apple size calculated its size influenced by the machine vision system, is relative. Aiming deficiencies of the prior art hierarchical conducted experimental studies presented in the diameter of the circumscribed circle of the average pixel classification methods.(3) The Apple absolute size fractionation techniques. As referred to herein refers to the absolute size of the calibration object after the introduction of Apple’s calculated based on the size of the pixel size of Apple, and Apple’s manual measurements are close to the absolute nature. Aiming deficiencies of the prior art hierarchical experiments were carried out, based on the proposed dynamic calibration of Apple absolute size fractionation methods.