An efficient and robust corner detection algorithm for furniture boards

In the intelligent inspection of furniture boards, wood debris generated during manufacturing can interfere with the imaging process. This leads to burr disturbances at the corners of the boards. Existing corner detection methods exhibit lower detection accuracy under the interference of these disturbances. To overcome this issue, this paper presents a corner detection algorithm tailored for furniture boards that incorporates the Random Sample Consensus (RANSAC)algorithm and line fitting techniques based on the Huber loss function. To enhance detection efficiency, our algorithm initially identifies horizontal and vertical edges near a corner. This preliminary step facilitates subsequent corner detection. This study introduces a test dataset for evaluating the accuracy and efficiency of algorithms. A battery of comparative experiments, including benchmarks with conventional methods, were conducted. The results demonstrate that our algorithm significantly enhances the efficiency and robustness of corner detection under a variety of complex burr interference conditions.