A LIBSVM quality assessment model for apple spoilage during storage based on hyperspectral data
Zhihao Wang1, Yong Yin1, Huichun Yu1
1College of Food and Bioengineering, Henan University of Science and Technology, Luoyang 471003, China. yinyong@haust.edu.cn.
Summary
This study developed a hyperspectral imaging spoilage benchmark and a LIBSVM model to assess apple quality during storage. The model achieved over 99% accuracy, demonstrating its effectiveness for long-term quality monitoring.
Area of Science:
- Agricultural Science
- Food Science
- Spectroscopy
Background:
- Assessing apple quality during storage is crucial for reducing food waste.
- Traditional methods are often destructive and time-consuming.
- Non-destructive techniques like hyperspectral imaging offer potential for real-time quality evaluation.
Purpose of the Study:
- To establish a spoilage benchmark for apples using hyperspectral data.
- To develop a robust quality assessment model for apples during storage.
- To validate the model's accuracy and applicability.
Main Methods:
- Hyperspectral imaging was used to collect spectral and image data.
- A spoilage benchmark was created using color, texture, and wavelet packet energy indicators.
- Successive Projection Algorithm (SPA) identified 20 feature wavelengths.
- Mahalanobis Distance (MD) verified the spoilage benchmark.
- LIBSVM model was trained and tested using pre-processed spectral data.
Main Results:
- The LIBSVM model achieved high accuracy: 99.94% (training) and 99.66% (test).
- Verification experiments showed 100% (training) and 99.83% (test) accuracy.
- The spoilage benchmark effectively correlated with apple quality indicators.
- The model demonstrated robustness and applicability for long-term storage monitoring.
Conclusions:
- The proposed spoilage benchmark and LIBSVM model are effective for assessing apple quality during storage.
- Hyperspectral imaging provides a reliable non-destructive method for quality control.
- The findings support the use of this technology in the food industry for quality assurance.
More Related Videos
Related Concept Videos
02:16Light Acquisition
In order to produce glucose, plants need to capture sufficient light energy. Many modern plants have evolved leaves specialized for light acquisition. Leaves can be only millimeters in width or tens of meters wide, depending on the environment. Due to competition for sunlight, evolution has driven the evolution of increasingly larger leaves and taller plants, to avoid shading by their neighbors with contaminant elaboration of root architecture and mechanisms to transport water and nutrients.
01:23Microbial Spoilage of Food
Microbial food spoilage refers to the degradation of food quality resulting from the metabolic activity of microorganisms such as bacteria, yeasts, and molds. These microbes proliferate on various food substrates depending on factors such as moisture content, nutrient availability, and storage conditions, leading to undesirable sensory and structural changes.Bacteria are primary agents of spoilage in high-moisture, nutrient-dense foods like meat, milk, and vegetables. Microbial spoilage occurs...