Enhanced Color Correction Techniques for Conveyor Belt Inspection Imagery via Unmanned Aerial Vehicles

This study addresses the challenge of color distortion in visual detection systems on unmanned airship-inspected underground belt conveyors at the Hongliu Coal Mine. Specifically, it tackles the issue of color information distortion under low-light or auxiliary lighting conditions. The research delves into the principles of the gray world color constancy algorithm and highlights the shortcomings of current methodologies. It introduces an enhanced gray world color correction algorithm that incorporates correlation and threshold weighting. This novel approach mitigates color information aberrations caused by low-light or auxiliary lighting and overcomes the single-color failure mode inherent in the original algorithm. Both simulated environment experiments and actual underground tests were conducted to objectively assess the algorithm’s performance. The findings demonstrate that the refined gray world algorithm effectively corrects color information in images under varied lighting conditions, exhibiting high robustness and fulfilling on-site operational demands.