Optimal Imaging Band Selection for Laser-Vision System Based on Welding Arc Spectrum Analysis

Abstract

The laser-vision system (LVS) is considered the most promising sensing method for robotic welding. Due to the harsh visual noise during the welding process, achieving accurate real-time detection of weld seams is still difficult. Since the optical propagation of the laser is affected by the welding environment, the commonly used visual noise filtering methods designed for the passive vision system (PVS) are no longer applicable. This article exploits the selection of the optimal imaging band for the LVS to eliminate the welding arc noise. The arc spectral distribution under various welding conditions is first measured, and five spectral bands with lower arc intensity are subsequently identified as candidates for the following investigations. Based on extensive investigations, we select the optimal imaging band for the LVS with the optimal visual clarity. Experiments have shown that LVSs working at this bandwidth can significantly reduce the interference of welding noise and improve imaging quality. Experiments show that the LVS working in this optimal bandwidth can effectively avoid the interference of welding noise, thereby achieving clearer visual imaging effects. To demonstrate the superiority of our findings for LVSs, we compare our method with a variety of well-known commercial products. This work reveals the arc spectral distribution under different welding conditions and provides effective guidance in enhancing sensing performance for LVSs.