Font Size: a A A

Research On Detection & Classification Of Egg's Dimension And Yolk Color By Machine Vision

Posted on:2002-03-10Degree:MasterType:Thesis
Country:ChinaCandidate:X W ChenFull Text:PDF
GTID:2168360122475449Subject:Agricultural mechanization project
Abstract/Summary:PDF Full Text Request
The duck's egg and its products, rich in protein, fat, vitancins ect, are very nutritious and have been a kind of important food in people's daily life. Especially, the salted egg, because of its peculiar taste ,is not only popular among out whole country but also a sort of traditional products to earn foreign currency. To produce salted eggs , it is necessary to classify eggs' dimensions and yolk color. Usually, egg processing plant detects eggs' dimensions by weighing then and detects yolk color by manual eyes. In this manual way, workers have to put eggs under light to expose the inside, then classify the grade by experience. The foal work conditions and the strong working intensity lead to low efficiency and plenty of detecting errors. So it is urgent to seek a scientific and practical method to detect eggs.According to the principle of manual means, it is possible to detect the yolk color and dimensions by machine vision. And such system is composed of software and hardware.The hardware system is consisted of CCD camera, video capture card, PCL-731A digital input/output card, signal delay controller, industry computer and outside executing set. When the egg-reached signal is sent to PC, through I/O card, the PC drives the video card to capture one frame picture to RAM. After a series of mathematical analysis, the PC send the yolk color's or egg size's grading signals to outer signal delayed controller by I/O card again. Last, the controller sends the delayed signals to outer execute equipment to finish classifying.The software is developped on the base of the hardware. And it has such functions as followings:(1) To process the image information of the control system.(2) To supply a GUI to operate the whole system.(3) To send classifying signals to outside.(4) To record the running situation.And this software was developed by C++ under Microsoft VC++ 6.0 and is composed of 5 modules: the initial module to finish the video card and I/O cardinitializations; the start-stop module to create and destroy the thread used to supervise light room and process image; the image-capturing module to capture the image from CCD to RAM; the image processing module to get classified information through analysis; the signal I/O module to send the signals to the outside and also to send outer information to the inside; the system log module to record the running situation of the system and the quantity of detected eggs.After abundant experiments, the system runs well. There are some conclusions drawn as followings:(1) All kinds of eggs, including cyan shell and white shell, there is a marked linear correlativity between the yolk color grade and three variable "H","S" and "I".(2) The yolk color grade lies on the "H" of the yolk region; the dimension of the egg has direct ration with the pixel area of the egg's picture.(3) Egg shell color is relate to the color information of the yolk region, which can be indentified by "I" by fuzzy mathematics.(4) The signal delay commission is done well by the controller.(5) The software could finish the egg dectecting and classifing well.
Keywords/Search Tags:egg classification, machine vision, controlling software
PDF Full Text Request
Related items