Test Similarity Data Ground Truth with Image Acquisition Using the Dice Algorithm

Abstract

In-vehicle tires there are essential components, one of it is a wire that supports the tire, there are several types of tire wire, one of it is brass plated tire steel cord. The object of brass plated steel cord tire has a micro-size below 1mm and a wave shape. In checking the quality of brass plated steel cord tire is usually measured manually by experienced experts, the manual measurement process sometimes experiences inaccuracies due to visual fatigue factors. Moreover, measurements must be repeated for the future so that it takes longer. The development of application technology using image processing is increasingly widespread, but the use of quality detection in steel cord and brass cord using image processing is are currently not found. The first step before taking an image measurement on brass plated tire steel cord is to test the similarity of the acquisition image with the data ground truth image on the brass plated steel cord tire so that it can be ascertained that the measurement stage can be done. The method used is related to the similarity test of this image using the similarity Dice algorithm. The results of the similarity test use Dice on 30 samples by comparing the data of Image Acquisition and Data Ground Truth. The truth brass plated tire cord has an average value of 99%. It can be concluded that the object edge detection of the image of brass plated tire steel cord has high accuracy for measuring using image processing techniques.