Automated Rust Detection via Digital Image Recognition during Grinding Work Process*

The rust removing process on the surface of a steel material by robotic systems is more efficient and effective than human’s work in the infrastructure maintenance. However, the existing rust removing robotic systems cannot automatically remove the rust according to the condition of the rusted area on the steel material. To solve this problem, several rust detection methods were proposed. Unfortunately, this problem is still a challenge due to the condition change during the rust removing process. In this paper, we propose a rust detection method which can work during the rust grinding process with a developed rust grinding platform. The proposed method is divided into two parts. The first part is called Qualified Image Detection (QID) to detect the rusted steel image without other objects included. The second part is the rust detection from the qualified image. Since the rust powder affects the rust detection result obviously, a technique is applied in our method to decrease the influence. The experiments were conducted on the developed grinding platform to show the validity of the proposed method.