Development and application of image analysis system for egg production factory

Egg production facilities have grown and production amounts have increased due to the developing technology and increasing demand. Quality and hygiene of egg production has come to the forefront. For the production of high quality and hygienic eggs, it is necessary to remove the defective eggs from the production line and the remaining eggs must be packed in accordance with the standards specified on the package. This process is called egg grading process. Within of the European Union standards, the Ministry of Agriculture and Forestry has issued a circular on the separation, classification and production of eggs produced and presented to the consumer. The facilities that produce eggs must produce within the limits determined in the circular published by the Ministry. Egg grading process can be carried out manpower, as well as with machines designed for egg grading. Depending on their production capacities, facilities prefer one of these two classification methods. Image processing is a technology that can be used to egg grading process. Using image processing techniques, it is possible to design fast and reliable systems compared to manpower, low cost systems according to existing grading machines. In this study, an image processing based embedded system design has been made in egg production facilities which separates the eggs coming from the conveyor belt to the desired classes. The images of the egg taken by the camera placed on the conveyor belt were evaluated by developed algorithms and the abnormal egg detection and dimensional classification of the egg were applied. In this study, Raspberry Pi microcomputer and Raspberry Pi the camera module were used. The program was developed using the python program with add-on the OpenCV library. Sensor integrated into conveyor belt, the rotation information of the belt has been obtained and it has been ensured that the image is taken from the camera periodically by using this information. Three images of each egg were taken so images of all surfaces were obtained. As a result of the images obtained, any abnormalities in the egg shell have been evaluated, and in the absence of an abnormality, dimensional analysis of the egg was made. Three different resolutions were used in application, the highest success was achieved in 1280x720 resolution. 120 eggs were used. Abnormalities were determined with 100% success and dimensional classification was achieved with a success rate of 92,85%.