Vision inspection system for halal chicken authentication

Abstract

Vision inspection system is a rapid, economic, consistent and objective inspection technique. This project focuses on the process to authenticate whether the slaughtered chickens images are halal and slaughtered correctly according to syariah law or not based on the slaughtered chicken images by using image processing methods and neural network. The captured images first undergo pre processing such as cropping, background subtraction process and erosion method. The pre processed images are then segmented using method such as component connected labeling method to detect the circularity of esophagus and trachea. Feature extraction is used to extract important information from the processed images and the features used in this project are circularity and the average value of red, green and blue (RGB) colours. Probabilistic neural network is used to classify the features of the image and compared with the trained feature to authenticate the halal status of the image. Time taken for authenticate slaughtered chicken image is fast and its accuracy is also reasonable.