同色异谱破坏特性及其在相似异物检测中的应用

由于相似异物与所处背景的外观特征极其相近，现有的光谱成像检测技术遇到新的挑战。针对复杂背景中相似异物检测难题，提出了利用同色异谱破坏特性的透反射光谱图像检测方法。在分析同色异谱破坏过程中异物与背景光谱特性差异基础上，建立了波长优选函数筛选最优波长的透反射光源构建成像检测系统。以棉花中的白色丝状异物作为实验对象，利用最佳波长透反射成像系统获取异物特征，采用图像决策层融合增强异物图像特征，最后利用二值化处理提取异物目标。实验结果表明，在利用390 nm反射波长与580 nm透射波长的双波长图像检测中，成像系统获取的相似异物图像特征明显，异物提取后检测结果与实际相符。此方法可有效检测棉花中多种白色丝状异物，为相似异物检测提供了一种新思路。

Metamerism Breakdown Characteristic and Its Application in Detection of Foreign Materials

Spectrum imaging detection techniques currently face the new challenge that it's difficult to detect similar foreign materials for their same appearance as the background.To resolve the problem of detection of similar foreign materials in the complex background,the objective of the present research was to develop a transmission/reflection spectrum image detection method based on metamerism breakdown characteristic.In the discrete spectrum wavelength region from 370 to 1 100 nm, this method is based on the principle that different materials have different spectral transmission and reflectance characteristic.When submitted to a source of illumination at different wavelength,foreign materials present different transmission and reflectance values in comparison to those from background.In the present paper,for simultaneously discriminating several types of foreign materials from background,the transmission and reflectance characteristic discrimination between foreign materials and background in the metamerism breakdown process was analyzed,then the optimal wavelength evaluation function for describing the background/foreign materials transmission and reflectance discrimination was set up to seek an optimal transmission/reflection illumination wavelength.Through the spectrum imaging experiment,the optimal operating condition including optimal detection wavelength and optimal illumination energy for these detected foreign materials was determined and accordingly an optimal wavelength imaging system was developed.In the actual application to the detection of foreign materials in cotton,the foreign material image features were acquired using the optimal wavelength transmission/reflection imaging system,also the image features were enhanced using image decision band fusion,then the targets of foreign materials were extracted effectively from image using binary processing.The experiment result indicated that the image features of similar foreign materials acquired by imaging system were obvious,and the detection result consisted with the fact in the double wavelength image detection system,which include 390 nm reflection wavelength and 580 nm transmission wavelength.This method provided an effective way for the detection of white fiber foreign materials in cotton and a novel practical means to solve the current problem of foreign material detection.