Supervisory System Based on Image Processing Using Library Open Computer Vision In Tool Management System
Gun Gun Maulana, Nur Wisma Nugraha, Silvia Ardya Garini · 2021
Currently, the supervisor of the availability of work tools in the tool management system is still done manually by the laboratory assistant in charge of recording the process of borrowing and returning work tools. However, there is already an automatic work tool detection system based on image processing. However, the system created often still has various errors in the detection results because there is often a decrease in the level of accuracy. This final project aims to improve the ability of open computer vision, which can replace human visual functions to support detection results to increase the accuracy of system performance with parameters of light intensity, camera distance, and work tool position. The development of the computer vision library is carried out using you only look once (YOLO) method to achieve an accurate tool management system and supported by room restrictions using automatic doors that can identify the users' identity. The result of the final project that has been made is that it can develop the performance of open computer vision to improve the accuracy of the detection system with an average accuracy value of 91%. The plan that has been made can be realized to record borrowing and returning transactions for work tools automatically and a website that can be used to provide information on the availability of work tools.