基于优化多光谱成像的异物检测新方法

为有效检测棉花中与棉纤维形态、色泽极其相似的异物,提出一种基于优化多光谱成像的异物检测新方法.在分析不同异物具有不同反射特性的基础上,在405~940 nm的波段范围内建立优化多光谱成像系统.通过实验确定了用于检测多种异物的最佳检测波长及最佳光照强度,在不同的光照条件下获取不同的光谱图像,然后采用小波变换图像融合算法获取完整的异物图像特征.实验结果表明,两种波长405 nm和850 nm构成的多光谱图像融合可检测多种棉花异物,异物检测结果与实际相符.此方法为解决棉花异物检测难题提供了一种新的途径.

A Novel Method for Detection of Foreign Materials Based on Optimal Multi-spectral Imaging

For effectively detecting foreign materials having the same shape and color as the cotton, the method based on optimal multi-spectral imaging is proposed. The objective of this research is to develop an optimal multi-spectral imaging system (OMIS) for detecting foreign materials in the spectral region 405～940 nm. This method is based on the principle that different materials have different spectral reflectance characteristics. Through experiments, an optimal wavelength and light energy are determined for detecting each certain kind of foreign material. Then, multispectral images of foreign materials are captured with a CCD camera at different optimal wavelengths source of illumination. An image fusion algorithm based on wavelet analysis is created to acquire complete information on foreign materials. Imaging results show that the combination wavelengths of 405 nm and 850 nm are the most appropriate for detection of a wide range of foreign materials, which provide an effective method for the detection of these foreign materials in cotton.