Introduction to Jujube Detection and Classification System Based on Machine Vision

China is the main producer of red dates. As an important economic crop in China, the requirements for quality inspection and classification are becoming higher and higher. At present, the traditional methods of detection and classification are as follows:

1) Manual classification: There are mature classification methods at home and abroad. Reverse fractionation and conveyor-type classification tables are used. These two methods are distinguished with the naked eye, which has the problems of low classification accuracy, visual fatigue, and low classification efficiency. Difficulties in the classification and export of red dates and other agricultural products.

2) Mechanized classification: large and small classifiers and weight classifiers, this method can only be classified according to the size and weight of agricultural products. The working principle of the size classifier is generally to judge the size of the fruit by means of holes, gaps, etc., and then classify according to the size of the horizontal diameter of the fruit. The weight classifier is based on the weight of the fruit. When applied to special fruits such as red dates, it cannot take into account its important characteristics such as surface cracks, scars, and mildew.

3) Photoelectric classification: At present, a general sorting method is used, which uses the absorption and reflection of light in different wavelength ranges on the surface of agricultural products to analyze its color characteristics, thereby determining the quality level of the fruit. According to the range of the light band, it can be divided into visible light detection and near-infrared detection. However, this method only judges the surface color characteristics of fruits and vegetables, and ignores other aspects. Therefore, it has a large one-sidedness, which is not suitable for the needs of comprehensive inspection of the quality of red dates or other fruits and vegetables.

    Now, after the new jujube detection and grading system based on machine vision has been proposed and successfully implemented, the disadvantages of the above schemes have been well overcome. As shown in the figure, its basic working principle is to continuously transport jujubes through a conveyor. And the speed of the transmission device can be adjusted by the inverter. When the red dates pass under the camera, the signal is detected by the sensor, which triggers the camera to perform image acquisition and image processing. Through the image preprocessing and pattern recognition, the classification signal of red dates is obtained, and then the classification signal is transmitted to the PLC, and the quality detection and classification of date is controlled by PLC.

Machine vision jujube detection and grading system

The industrial cameras, industrial lenses, and industrial light source solutions used in the solution are provided by iJUNCO, and are carefully constructed using the technical experience accumulated by iJUNCO for many years. They have the top level of domestic machine vision software and hardware and are well matched In order to meet the needs of this solution, we will show its actual results through case pictures below.

Machine vision red date shooting hardware solution

Gigabit Ethernet interface 120WCCD camera + ring light source shooting effect

Machine vision red date shooting hardware solution

Gigabit network interface 120WCCD camera + diffuse reflection light source shooting effect

In addition to the use in jujube quality inspection and sorting, in modern industrial automation production, machine vision technology has universal applications in product inspection, production process monitoring, and part identification, such as inspection of the size of parts and components, and automatic assembly line parts. Integrity check, automatic positioning of electronic components, character recognition of integrated circuits, etc. Usually, the human eye cannot perform the highly repeatable work continuously and stably for a long time, and must have a certain intelligent work. When other physical quantity sensors are difficult to perform, machine vision technology can make a big difference.

The application of machine vision technology in agriculture mainly includes agricultural robots, agricultural remote sensing analysis, food and fruit quality inspection, etc. Machine vision has obvious advantages in fruit quality inspection and grading: it can ensure the consistency of inspection results and ensure the accuracy of inspection, especially in the calculation of surface defect area and coloring area. Its advantages are beyond human ability. Therefore, the application of machine vision technology to fruit quality inspection will greatly reduce the labor input, which will have very realistic social and economic significance for reducing the production cost of fruit commodities and improving its market competitiveness. On the other hand, the application of machine vision technology in fruit quality inspection has opened up new research fields for machine vision technology research, enriched the connotation of machine vision technology, and has high value.