Automated Book Defect Detection for Quality Inspection in China’s Publishing Logistics Using LME-YOLO

China’s publishing industry continues to expand across production, warehousing, circulation, and export. Book quality inspection affects product grading, logistics sorting, customer satisfaction, and the international image of Chinese publishing. However, these defects are often small, irregular, and low-contrast, and their visual similarity to paper textures and printed content complicates automated classification and localization. To address this problem, this study proposes Local-detail Modelling and Edge-enhancement YOLO (LME-YOLO), a book defect detection model based on YOLO11n for quality inspection in China’s publishing logistics. On a dataset of 632 book images, LME-YOLO achieved a Precision of 0.829, Recall of 0.777, mAP50 of 0.818, and mAP50–95 of 0.465. The model contains 2.64 M parameters and requires 6.6 GFLOPs. Among the detectors compared in this study, LME-YOLO achieved the highest mAP50 and mAP50–95. Compared with YOLO11n, its mAP50 and mAP50–95 increased by 0.015 and 0.006, respectively, with increases of 0.05 M parameters and 0.2 GFLOPs. Compared with YOLOv8, its mAP50 and mAP50–95 increased by 0.017 and 0.005, respectively, while the parameter count and computational cost decreased by 0.37 M and 1.6 GFLOPs.