The Improvement Of Duck Egg Quality's Nondestructive Automatic Detection & Grade System By Machine Vision

The duck egg, which is high in folks favor , contains high protein, low fat, kinds of vitamins and is a kind of animality food. In order to improve the duck egg's quality and production efficiency, reduce the labor intensity. accelerate the development of birds industry, it is necessary & impending to research on the nondestructive & prompt automatic detection technology.This research is the prophase's continue & thorough. The system uses the machine vision technique, multi-thread and digital image process technique to detect the duck egg. First, according to the egg's weight, egg's color, the depth of greenness and the thickness of eggshell, we set up the math-model between the image process parameters and determination targets Then using the computer vision technique, the CCD collects one complete dynamic duck egg's image to computer EMS memory, which will be processed. In the period of digital image process, the program uses the automatic searching arithmetic to one definite image and greatly reduces the time of image process, which is the base of multi-channel working system at the same time. By using multi-thread technique the system realizes the multi-channel working at the same time and makes the most use of the hardware. So it improves the efficiency and proportion of capability and price. The system makes most use of the efficient PLC, makes it satisfy the system necessity and makes it become the bridge of PC and execute machine.The system is composited by CCD, image collection card. PC. digital I/O card, PLC and execute machine. While the duck egg is coming into the darkroom, a signal which is triggered by one trigger circuit will be sent to PC through the PCB and notices PC to collect one complete duck egg's image to computer EMS memory, and then PC automatically processes the image and uses the math-model to get the targets which can classify duck eggs. In the end PC sends the classification signal to PLC to delay a definite time and then classify the duck eggs.In the period of synthetical testing, the color classification precision of the error in 1 grade is 83%. the error in 2 grade is 91%, in 3 grade is 97%. The weight classification precision of the error in 1g is 81%, the error in 1.5g is 86%, and in 3g is 98%. The depth of greenness precision is 82%. The thickness of eggshell precision is 81%. The precision of these parameters has met the factory's demands. On a whole the system can run steadily and won't appear the phenomena of invalid operation and computer crash. The interface is friendly, convenience and prompt. Now the sample machine has been used, and in order to find the further bug the system must be tested for a longer time.