|
|
自动化学报 2008
Robust and Efficient Cotton Contamination Detection Method Based on HSI Color Space
|
Abstract:
Efficient detection and separation of cotton contamination is important for controlling the quality of cotton. A robust and efficient method of detecting the contamination in cotton is proposed. In this method, 3D-LUT is used to reduce the time of transformation between different color spaces. Cotton contamination is detected based on the discrimination of hue and saturation in HSI color space between cotton fibers and cotton contamination. Color motion compensation is used to improve the spatial correction and reduce the wrong recognition rate. The experiments show that our system can complete the total process of image capturing, format transformation and impurities identification in 10ms. Under a velocity of the cotton at 18m/s, the recognition rate can reach as high as 95.4% with an impurity size of 2×2mm2.
