Computer Vision System For Monitoring Body Discomfort In Manufacturing Environment

• Main Author: Ramdan, Nur Sufiah Akmala
• Format: Thesis
• Language: English; English
• Published: 2016
• Subjects: T Technology (General); TA Engineering (General). Civil engineering (General)
• Online Access: http://eprints.utem.edu.my/id/eprint/18359/1/Computer%20Vision%20System%20For%20Monitoring%20Body%20Discomfort%20In%20Manufacturing%20Environment.pdf; http://eprints.utem.edu.my/id/eprint/18359/2/Computer%20Vision%20System%20For%20Monitoring%20Body%20Discomfort%20In%20Manufacturing%20Environment.pdf; http://eprints.utem.edu.my/id/eprint/18359/; https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=100172

Manual handling is one of the primary causes of body discomfort. If motions are repeated frequently, such as every few seconds, and for prolonged periods such as an eight-hour shift, fatigue and muscle strain can happen. Body discomfort can occur in every part of the body, such as the neck, arms, waist, spine, legs, and feet. To reduce body discomfort among manual workers, many researchers have carried out studies on body discomfort. Previous studies usually used the traditional method, which is by carrying out surveys of manual workers using specific questionnaires. A questionnaire that is often used is the Nordic Musculoskeletal Questionnaire (NMQ). The questionnaire is designed to find out about discomfort that occurs in all parts of the subject’s body. The traditional method cannot detect body discomfort automatically because no automatic system is used. Furthermore, the Closed-Circuit Television (CCTV) used in factories nowadays is used for security purposes, not for ergonomic purposes. Therefore, the goal of this research is to design a vision system that monitors body discomfort in manual workers. It is done by using a new method, the image histogram. The methodology proposed is the development of a vision system using Python and SimpleCV software for recording images and image analysis. The output of the image analysis is a red-green-blue (RGB) histogram which shows the pixels of the gray scale color distribution. The image analysis is done every three minutes for 30 minutes. The results show that when the worker is moving in order to carry out his or her work, the RGB histogram also changes. When the histogram is changing throughout the period of 30 minutes, it is found that the person is likely to feel body discomfort regardless of which part of the body is involved. By also referring to the image frame, it is proven that the worker is experiencing body discomfort within the 30 minutes. To support and strengthen the results, NMQ analysis is also used. The experiments are done by conducting three types of experiments on the fitting process, milling process, and turning process. Nine subjects participated in these experiments. The results show that seven of the subjects experienced body discomfort in the range of the hypothesized limit time. From all the results, including the histograms, it is shown that the system can monitor body discomfort successfully.