Research on nondestructive detection of pine nut quality based on terahertz imaging

•In this paper, a rapid and nondestructive detection of pine nuts for mildew and plumpness based on terahertz transmission imaging technology was carried out, and four preprocessing algorithms of AirPLS, ALS,BEADS and SNV + Detrending are used to remove the redundant information and noise.•The LDA, SVM and XGBoost integration learning discriminant models were established respectively to explore the optimal qualitative discriminant model for the detection of pine nut quality, and the research shows that the BEADS + XGBoost qualitative discriminant model has a recognition rate of up to 98.61%.•The main feature gamut extraction and channel separation preprocessing were carried out for redundant color information of the 0.1–2.0 terahertz pine seed images, and the appropriate threshold is calculated by the one-dimensional maximum entropy threshold segmentation algorithm with the error less than 7%.•This study provides a new rapid and nondestructive effective method for pine nut quality detection, which can provide technical reference for other shelled nut quality detection and has significant practicalvalue.