Development of a computer vision system for ensuring quality of Fetal Spiral Electrodes

Abstract

Computer vision systems are becoming increasingly common in industrial applications. One of the areas where they are being used more and more is in quality control, as they can provide both increased efficiency and effectiveness. This thesis aims to develop a computer vision system that can quality control the dimensions of a fetal spiral electrode. The electrode is used to monitor the fetal heart rate during childbirth, by inserting the spiral electrode into the scalp of the fetus. The wrong dimensions can cause the creation of a wound on the baby’s head, and thereby potentially cause an infection. This new computer vision system is needed because the current manual measuring solution lacks precision and there is a lot of room for human errors to occur, which can affect the measuring results. Four different parts were conducted, forming a project consisting of knowledge acquisition, concept development, system development, and evaluation of the system. These resulted in a prototype performing height measurement and tip angle measurement of the spiral. This is done by binarizing the image by thresholding and then processing and evaluating the contour to find important points, shapes, and changes in the image. It was also done by creating a fixture that holds the electrode in position and optimizes the external conditions. The evaluation shows that the developed computer vision system performs with higher precision and accuracy compared to the currently used system.