The Research Of Unfertilized Eggs Automatic Verification System Based On The ARM Embedded Platform

With the development of livestock husbandry, advanced science and technology has changed the traditional form of animal husbandry and production operations. Hatching production by microcomputer control incubator has replaced the old incubator.However, the way we eliminate unfertilized eggs before hatching is still using handheld simple as mixer.The workload of this method is big, slow and human factors influence is bigger. Moreover only on the fifth day of hatching head as can we eliminate the unfertilized eggs.While fresh eggs will drop below for series b after being hatched for 72h,unfavorablly processed into egg products,causing useless energy.Besides plenty of eggs are wasted and the bacteria of those eggs which are unable to be hatched out will pollute normal eggs. Breakage is higher; especially when eggs turned to stay longer this affects the increase in hatching rates of eggs. This paper adopts advanced machine vision technology, combines the technology of digital image processing with embedded technology, puts forward the design scheme of unfertilized eggs verification system based on embedded computer.This system making use of the images of hatching eggs is collected by using computer vision technology, and then, edge character parameters of image area is extracted. At last, identification of unfertilized eggs based on trained genetic algorithm neural network, can improve the speed and accuracy of eliminating and reduce the damage and pollution. It has certain practical significance to improving hatching rates. This paper also provides a new thinking and technology with eliminating unfertilized eggs.The main research contents of this study and the results are as follows:1,We summarized the detection research progress and status of machine vision for hatching eggs at home and abroad, pointed problems existing in similar researches of domestic and foreign, researches based on machine vision detecting method, and urgent needs for hatching eggs detection equipment development.2,We established a system based on ARM embedded platform for this study of unfertilized eggs verification. This system hardware component include the ARM S3C2410 embedded system developing board and CMOS image sensors OV7620, put them through interfaces connected together constitute a basic embedded unfertilized eggs verification system. According to the system demand, in the ARM S3C2410 development platform by choosing the support of ARM9 HY57V561620CT - H chip (SDRAM), K9F1208 chip (FLASH) and two LED lights and a buzzer.3,We collected the image data of hatching eggs, via grayscale image initialization, median filtering, and Newton dichotomy extracting eggs. With contour following algorithm edge we extracted invariant parameters(circularity,complexity,elongation,globularity,ratio of major to minor axis and alteration coefficient).At last we used trained genetic neural network to release unfertilized eggs verification technology. The test result shows: this system is convenient and easy to realize. The identification rate of 600 non-samples of the hatching eggs as verification samples is 99.3 percent, which can absolutely satisfy the require of eliminating the unfertilized eggs before hatching.