Microscopic hyperspectral imaging technology is a potential non‐destructive and non‐contact method for colour measurement of micrometre‐sized textile fibres. However, specularity on the fibre surface can distort the accurate colour information and affect the accuracy of the colour measurement. This paper proposed a specular‐constrained sparse approximation (SCSA) for specular‐diffuse reflection separation from hyperspectral images of wool fibres. First, a specular prior map is generated based on the lightness dissimilarity. Then the SCSA model is used to decompose the processed hyperspectral image A into low‐rank data L, sparse specularity data S constrained by the specular prior map, sparse noise E, and Gaussian noise N. A non‐linear logistic sigmoid function and a sparse approximation of A – L – N to S are used to improve the performance of specularity removal during iterative optimization. The experimental results show that the proposed method significantly preserves diffuse reflectance and texture details in the specular highlight regions to obtain actual spectral reflectance and chromatic values from hyperspectral images of wool fibres.
Prior knowledge of textile fiber colors in blends is useful for color recipe assessment. There have been various methods to improve the accuracy of color recognition from the dataset of colored textile fibers in recent years. However, numerical assessments based on spectral feature and color difference is insufficient, in which the accuracy of color recognition can be affected by morphology and the uneven coloration on a single fiber. This paper proposes a novel 3D convolutional neural network model (3D‐CNN) with supervised spectral regression for the color recognition of hyperspectral images (HSI) of colored textile fiber. The proposed method obtained spatial‐spectral features based on 3D‐CNN, and the true spectrum of each class was used for supervised spectral regression to improve the accuracy. The loss function used was the sum of the supervised classification loss function and the spectral regression loss function model are optimized by mini‐batch‐based backpropagation. The proposed method was trained and tested on the HSI dataset composed of 100 colors of wool fibers acquired through a microscopic hyperspectral imaging system at a ×3.375 optical magnification. The experimental results showed that the proposed method exhibited better performance compared to numerical assessments and other deep learning models, except for efficiency. Specifically, it achieved better recognition performance on sub‐datasets of similar colored and light‐colored wool fiber where subtle inter‐class and large intra‐class variance existed.
In order to make a color assessment for micron‐grade materials such as a single fiber, a non‐destructive and push‐broom microscopic hyperspectral imaging system (MHIS) was set up consisting of a stereomicroscope, an imaging spectrograph, and a digital detector. The performance of this system for acquiring accurate and repeatable information, such as spectral and colorimetric values, was investigated. The experimental results show that the system has an excellent spatial resolution, repeatability, and accuracy. The spatial resolution with a ×1.5 auxiliary objective lens was down to 90.244 lp mm−1 (5.54 μm). The average ΔE00 of the tested patches in the ColorChecker Mini chart was between 0.15 and 0.40, and the color measurement repeatability was acceptable. Color difference ΔE00 and the standard variation of spectral wavelength between ×2.5 and ×4.5 magnification were larger than those at lower magnifications. It can be concluded that the MHIS can identify colorimetric values of the materials and detect the color changes sensitively with a high spatial resolution at micron‐grade.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.