Monitoring and Alarming Activity of Islamic Prayer (Salat) Posture Using Image Processing

This paper introduced a Salat Inspection and Training System based on Machine Vision and Image Processing Subject. In Islam, prayer (i.e., Salat) is the second pillar of Islam. It is the most important and fundamental worshipping activity that believers have to perform five times a day. From gestures’ perspective, there are predefined human postures that must be performed in a precise manner. There are lots of materials on the Internet and social media for training and correction purposes. However, some people do not perform these postures correctly due to being new to salat or even having learned prayers incorrectly. Furthermore, the time spent in each posture has to be balanced. To address these issues, we propose to develop an assistive intelligence framework that guides worshippers to evaluate the correctness of their prayer’s postures. Many features of images are being extracted and analyzed. Methods for image comparison and pattern matching are used to study the system’s effectiveness by using several combining algorithms, such as Euclidean Distance, Template Matching and Grey-Level Correlation, to compare the images of the user and the database. The experiments’ results, both correct and incorrect salat performances, are shown via pictures and graph for each of the postures of salat. Limitations of the system, such as lighting, is discussed regarding how it affects the system’s performance.