Applications of computer vision techniques to cotton foreign matter inspection: A review

• Applications of computer vision to cotton foreign matter inspection are reviewed.
• The use and improvements of image processing algorithms for cotton foreign matter are introduced.
• Six types of computer vision systems for the inspection of cotton foreign matter are reviewed.
• The potentialities of computer vision for the inspection of cotton foreign matter are explored.

Cotton is an important crop throughout the world, and its quality plays a significant role in its profitability and marketability. Foreign matter in cotton can cause damage to spinning, weaving, and dyeing and thus seriously affects the quality of cotton products. Conventional methods including inspection by human workers and instrument based approaches such as photoelectric detection and ultrasonic detection are time-consuming, labor-intensive, and sometimes inaccurate. As a non-destructive, cost-effective, rapid, and objective inspection tool, computer vision has been widely used in cotton foreign matter inspection. In this review, the basic concepts, components, and image acquisition modes of computer vision techniques are presented. The improvements in image processing and analysis of foreign matter in cotton are introduced, and several different computer vision systems that have been created to detect foreign matter are reviewed to highlight the potential for the inspection of foreign matter. Considering the progress made to solve this type of problem, we also suggest some directions for future research.