Detection of fake shallots using website-based haar-like features algorithm

Abstract

Shallots is commonly used as essential cooking spices or complement seasoning. The high market demand for this commodity has triggered some people to counterfeit it. They mix the shallots with defective products of onions to get more benefits. It urges to provide a system that can help people to distinguish whether the shallot is original or fake. This research aims to provides an object recognition system for fake shallots utilizing the Haar-Like Feature algorithm. It used the cascade training data set of 59 positive images and 150 negative images with 50 comparison images. The identification process of the shallots was through the haar-cascade process, integrated image, adaptive boosting, cascade classifier, and local binary pattern histogram. This system was made based on the Django website using the python programming language. The test was conducted 30 times on Brebes shallots mixed with Mumbai's mini onions in a single and mixture test method. The test obtained an average percentage of 69.2% for the object recognition of Mumbai's mini onions.