BUTTER QUALITY CONTROL INSPECTION USING A COMPUTER  VISION APPROACH

Henok, Gashu

dc.contributor.author	Henok, Gashu
dc.date.accessioned	2022-11-16T11:06:46Z
dc.date.available	2022-11-16T11:06:46Z
dc.date.issued	2022-04-21
dc.identifier.uri	http://ir.bdu.edu.et/handle/123456789/14393
dc.description.abstract	Agriculture is one of the most important sources of food. The quality of food is influenced by the source. In response to increasing health issues and customer needs about safe food options, demand for high-quality food products is rising globally at unprecedented rates. Butter is a common food item produced from cow's milk that is high in nutritional value. There are several methods to control the quality of agricultural food production. Among these techniques, Machine vision is a new technology for acquiring and investigating an image of a natural scene by computers to control machines or process it. Agricultural products are evaluated using machine learning. Butter can be applied in day-to-day edibles. Pure Butter is the most sources of nutrients and fats. However, Butter is available on the market within other adulteration mixed items, like, banana, buttery, potato, gully, fat, and milk powder or flour with milk. Butter's quality inspection is not easy because the impurity and the Butter itself have similar colours and features. We prepared our dataset using food processing engineering laboratory mixer in the ratio of 80:20 and 90:10 for training and performance evaluation. The major goal of this research is to generate a computer vision model that can detect butter quality. To achieve the study's aim, we collected pure butter and prepared our dataset, captured using a smartphone. After the image acquisition, we applied pre-processing to reduce noise. The preprocessed image input for feature extraction using CNN and GLCM. We used CNN and KNN data classifiers. Here we considered combined approaches of handcrafted and automatic feature extraction techniques. Finally, quality inspection using end-to-end deep learning techniques and KNN is applied. From the experimental results, we registered an accuracy of 81.75%, 95% and 90.5, xi 93% in both ratios 90:10 and 80:20 fresh & not fresh respectively. To improve the result, we have used principle component analasis (PCA). Keyword: Computer Vision, Butter, Fresh, Pure butter, Mixed butter, Food, Process, Quality, Control Inspection, Machine Learning, K-NN, CNN, GLCM, Milk, Image classification, colour.	en_US
dc.language.iso	en_US	en_US
dc.subject	FACULTY OF COMPUTING	en_US
dc.title	BUTTER QUALITY CONTROL INSPECTION USING A COMPUTER VISION APPROACH	en_US
dc.type	Thesis	en_US