Automatic Tracking Camera System for Construction Machines by Combined Image Processing

Abstract

The damaged area caused by disasters must be restored as soon as possible. So, remotely controlled construction machines are used to execute this dangerous work, safely. In the remote construction, machine operators always watch images of the working field from pan-tilt monitoring cameras controlled by camera operators. Then, the automatic camera control system without camera operators is expected for machine operators. Based on the above requirement, we are developing an automatic tracking camera control system for construction machine operators by using image processing. The most difficult point for this system is to realize a stable outdoor object detection by image processing. So, we proposed a new image processing method which is a combination of "Improved template matching" and "Moving object detection". From experimental results, we confirmed that the proposed method is improved about the robustness of outside construction machines tracking.