Color Measurements Based On Hyperspectral Imaging Method

Color is an important basis for determining product quality and identifying object types.Current color measurement methods such as color cards,color cameras,spectrophotometers,and the like have certain limitations in measurement accuracy and efficiency: the method of manual comparison using color cards is inefficient and accuracy is limited;based on red green blue(RGB)principle color camera imaging efficiency,but its spectral resolution is relatively rough,can not distinguish the color of the tiny color difference;spectrophotometer has a good color measurement accuracy,but it can only be a single point of measurement,color measurement efficiency is very low.Hyperspectral imaging technology is a brand-new remote sensing imaging technology developed in the past decade.It has features such as super-multiband,high-spectral resolution,narrow band,wide spectral range,and map integration.It is widely used in geology,vegetation science,agriculture and other fields.This topic applies hyperspectral to the field of color measurement and designs a hyperspectral based color measurement method.Compared with the existing colorimetry methods,the hyperspectral colorimetry method can measure a wider spectral range and can record more finely.The spectral features have a higher color measurement accuracy.In order to solve the difficult problem of distinguishing two similar colors,this project designed and built a prismatic spectrophotometric method based on the principle of hyperspectral imaging technology.Complementary Metal Oxide Semiconductor(CMOS)as Focal Plane Array(FPA)The hyperspectral imaging system of the material is applied in the field of color measurement,and the imaging spectrum of the international standard color card is verified,which verifies the feasibility of using the hyperspectral imaging system to measure the color.In addition,combined with the specificity of hyperspectral image data,a series of data processing algorithms such as non-uniformity correction,pixel binning,band combining,and radiation normalization have been specifically designed to obtain rich hyperspectral data obtained through experiments.The color information collected by the color camera was quantitatively compared by using the subdivision band comparison,the combined band comparison,and the spectrum angle matching comparison,which proved that the hyperspectral image has a wider spectral range and finer spectral characteristics.Can distinguish the color difference that the color camera can't distinguish,the effect is very remarkable,it is of great significance to improve the color measurement accuracy in the industrial production.