Camera Placement Optimization for Vision-based Monitoring on Construction Sites

For successful vision-based monitoring, the camera placement is an important and challenging issue. However, the previous Art Gallery Problem – which is a traditional optimization problem closely similar to the camera placement problem – did not take into account characteristics of construction sites, and thus faced difficulties in placing cameras on site and collecting adequate video/image data of construction activities. To handle the drawback, this paper aims to develop a camera placement optimization framework that considers characteristics of construction sites. The framework consists of two main processes: (1) problem definition and mathematical modeling and (2) optimum camera placement. For the validation, the case study was performed. The results showed the applicability of the developed framework in finding the optimal number, locations, and viewpoints of surveillance cameras for vision-based site monitoring. In conclusion, this research can contribute to propose proper camera placement on construction sites, increase quality of collected visual data, and improve the performance of site monitoring.