Plastic mulch used in cotton cultivation tends to age and fragment, and the resulting pieces can blend with cotton fibers during harvesting, forming transparent heterogeneous fibers that severely degrade lint and textile quality. Effective identification of such impurities has thus become a priority in the textile industry. A Mueller matrix measurement platform was constructed to characterize carded cotton, uncarded cotton, and plastic mulch samples under both transmission and reflection modes at four wavelengths (520, 532, 638, and 650 nm). Polarization features were extracted, and three classification strategies were proposed and experimentally validated. Results show that wavelength variation most strongly affects the matrix elements of carded cotton, followed by uncarded cotton. The matrix elements of the plastic mulch samples show no significant variation with increasing wavelength. An increase in the reflection angle enhanced the ability of the cotton samples to modulate the incident light in the horizontal and vertical directions. It concurrently increases their depolarization effect, whereas the residual mulch shows decreased phase retardation. The group tests were conducted using three classification strategies and two classifiers (LDA and KNN). The optimal solution was achieved by combining the hierarchical classification strategy based on matrix features with the KNN classifier under transmission mode, yielding a 95 % classification success rate.This research focuses on the intrinsic polarization properties of cotton plastic mulch, which differ from the traditional image-based identification principles. This approach not only enriches the understanding of the polarization features of cotton plastic mulch but also provides a new research foundation and theoretical support for future advances in cotton processing technologies, such as impurity detection and online recognition of plastic mulch. • Measured Mueller matrices of cotton and plastic film under varied conditions using a polarization system. • Cotton exhibits significantly greater polarization sensitivity to wavelength variation compared to plastic film. • The carding process altered the polarization characteristics of cotton. • A hierarchical classification strategy using matrix features and a KNN classifier achieved 95 % accuracy.
J et al. (Mon,) studied this question.