A Rust Removed Region Detection for Automated Rust Detection during Grinding Work Process

In the maintenance task of steel-made infrastructure, such as bridges and towers, the rust removing process is one of the most dangerous and hard works for human beings. The robotic systems can provide a more efficient way to accomplish the tasks. However, the existing robotic systems cannot automatically remove the rust according to the rusted area condition during the rust removing process. As part of the solution, several rust detection methods were proposed. Although they can inspect the rust from a given image, the condition changes during the rust removing process affect the result obviously, which makes them difficult to be utilized in the real application. In our previous research, a rust detection method utilizing a sequence of digital images was proposed by observing the rusted area conditions during the rust grinding process. However, the ROI (Region of Interests) need to be manually set to estimate the change of the rusted area in the previous method. In this paper, we propose a method to determine the cleaned rusted region by an image processing technique with a low pass signal filter. The experiments were conducted on the developed rust grinding platform to show the validity of the proposed method.