Abstract: With the development of machine vision systems, auto-sorting systems have been used for removal of foreign fibers (contaminants) in cotton. Of them, white foreign fibers are hardly distinguished from lint cotton because their colors are the same or very close under the illumination of visible lights and ultraviolet lights in the inspection by an existing machine vision system or manual sorting. The laser imaging methods previously suggested by our team can detect most of the white foreign fibers in cotton, but there are 2 problems that need to be solved: One is that they can only use area cameras for the inspection, which is not suitable for the processing line of the lint cotton in textile industry; the other is that the optimization of parameters for the laser imaging system is only concerned with the camera exposure time, but not concerned with the wavelength and the power of the line laser. To solve the 2 problems, in this paper, 12 kinds of typical non-fluorescent white foreign fibers random distributed on the cotton surface were driven by a conveyor, and 300 frames of their images were obtained by a high-speed line-scan camera under the illuminations of line laser at the wavelengths of 658, 405 and 850 mm with different laser light powers and different exposure time individually. Firstly, the contrast of the white foreign fibers and cotton background in a frame of the laser image was quantized by the difference of their average pixel values. Then, the relationship curves between the quantized differences of fiber and cotton grey values and the laser light's wavelengths, the laser light's powers and the exposure time were plotted and analyzed. It was found that using the optimized imaging parameters, the gray level of the foreign fibers had reached a saturated status, but the gray value of cotton was still unsaturated. Thus, by the significant difference of the targets and background, the white foreign fibers could be separated from the cotton. In the experimental imaging system, the image contrast of 12 kinds of white foreign fibers and cotton background had reached a maximum value when the laser images were obtained under the illumination of the line laser at the wavelength of 650 nm with the power of 55 mW and the exposure time of 36 μs. Finally, using the optimized laser imaging system and the traditional imaging system of LED plus ultraviolet light individually, 730 frames of the images of 12 kinds of white foreign fibers with lint cotton were obtained. The experimental results indicated that, the 12 kinds of white foreign fibers could be easily distinguished from cotton. With the optimized imaging system, using a simple algorithm of Prewitt edge detecting or fixed binary segmentation, the successful detecting rate was up to 93.7% or 92.9%, respectively.