In-line semantic segmentation of kimchi cabbage deterioration using YOLOv8n and DeepLabv3

• Published in: Postharvest biology and technology Vol. 218; p. 113158
• Main Authors: Yang, Hae-Il, Min, Sung-Gi, Yang, Ji-Hee, Eun, Jong-Bang, Chung, Young-Bae
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.12.2024
• Subjects: Agricultural quality control; Convolutional neural networks; Edge devices; Plant disease detection; Real-time monitoring; BG; FN; FP; HT; ASPP; IoU; Real-time monitoring; SD; Plant disease detection; AFPS; MD; KC; CVAT; Agricultural quality control; Edge devices; TP; Convolutional neural networks
• ISSN: 0925-5214
• DOI: 10.1016/j.postharvbio.2024.113158

To ensure the safety of kimchi cabbage (KC), the effective identification of deterioration is required. In this study, we aimed to develop and validate a semantic segmentation pipeline to detect and segment the mild and severe deterioration of KC on moving industrial conveyor belts. The pipeline integrates the YOLOv8n and DeepLabv3+ models, image processing and semantic labelling models, respectively. The YOLOv8n model was employed for initial detection, achieving an F1 score (a measure of precision and recall) of 0.988 and an AP50 (average precision at 50 % intersection over union (IoU) with the ground truth) of 0.994. For detailed segmentation, the DeepLabv3+ model with Resnet101 and Mobilenetv3 backbones was used, and the performance of these models was evaluated based on F1 scores. The YOLOv8n model demonstrated high accuracy in detection. Further, the DeepLabv3+ model with the Resnet101 backbone achieved F1 scores of 0.772 and 0.851, and IoUs of 0.628 and 0.741 for mild and severe deterioration, respectively. Notably, the DeepLabv3+ model with Mobilenetv3 backbone excelled in identifying severe deterioration, indicating its potential for generating rapid alerts in edge devices. Crucially, the developed methodology enables the real-time, in-line monitoring of KC deterioration and is adaptable for future edge device applications. The results of this study highlight the potential of accessible technologies in advancing agricultural quality control, setting a benchmark for future research to expand data sources and enhance detection capabilities. •Semantic segmentation pipeline using YOLOv8n and DeepLabv3+ for quality control.•High detection accuracy achieved with AP50 of 0.994 using YOLOv8n.•Segmentation F1 scores enhanced using DeepLabv3+, up to 0.906.•Benchmark for real-time monitoring for detecting kimchi cabbage deterioration.•Demonstration of accessible technologies to improve agricultural quality control.