Analysis of the effectiveness of natural treatments for preserving apricots and the YOLOv7 application for early damage detection

Maintaining the quality of apricot fruits during storage is not an easy task due to the changes in their physical and chemical properties, so it is necessary to use less expensive, easy to apply, environmentally friendly, and safer preservatives to maintain the nutritional value of apricot. The damage to some fruits during storage can be a source of infection, which leads to the damage of healthy fruits more quickly, which requires building an intelligent model to detect damaged fruits. The aim of the research is to study the effect of immersing apricots in lemon juice once and sugar-water solution again on the quality properties of apricots, including sweetness, color, hardness, and water content. On the other hand, the YOLOv7 algorithm was used to detect healthy fruits and damaged areas using a camera. The results showed that sweetness increased with increasing immersion time in sugar-water solution to reach 22.1 Brix, while it decreased with increasing immersion time in lemon juice to 19.12 Brix. Also, hardness increased with increasing immersion time in sugar-water solution to reach 3.7 kg/cm2. The water content of apricots decreased with increasing immersion time in different immersion media from 77.14 g to 73.93 g. In addition, CIE-L*a*b levels increased with increasing immersion time in different immersion media. For the performance indicators of the YOLOv7 algorithm, precision of 84.5%, recall of 87%, F1 of 0.77, and mAP@0.5 of 77.2 were obtained, respectively. Therefore, this study is expected to reduce the workload in post-harvest fruit processing and help in the rapid identification and detection of damaged fruits based on smart detection algorithms, thus improving sorting efficiency and reducing both waste and economic losses, which enhances smart agriculture technologies.

Text 120426.pdf - Published Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (2MB) Official URL or Download Paper: https://link.springer.com/article/10.1007/s44187-0...

Actions (login required)

View Item