Early Warning Potential of Banana Spoilage Based on 3D Fluorescence Data of Storage Room Gas

In order to realize the early warning of spoilage during banana storage, 3D fluorescence data of the storage room gas corresponding to two batches of bananas with different storage dates were collected. On the basis of scattering elimination and Savitzky-Golay (SG) smoothing treatment of these collected fluorescence data, three methods, namely, contour maps, fluorescence regional integration (FRI), and systematic cluster analysis (SCA), were employed to define the benchmark of banana spoilage. Then, to eliminate the dimension influence between different physical and chemical indicators, principal component analysis (PCA) was used to fusion the physical and chemical data of different indicators, and the principal component was selected as fusion data. And then, according to least square regression (LSR), the feature wavelengths were selected for the two batches of test data, and the same and similar ones as the common feature wavelengths were also picked. Finally, a unified spoilage benchmark for the two batches of bananas was constructed by the common feature wavelengths. Based on their respective benchmarks and the unified benchmark, Mahalanobis distance (MD) was adopted to realize the early warning of banana spoilage with two batches. The research results show that the MD can realize early warning of banana spoilage during storage for different batches according to the unified benchmark. This shows that the determination methods of the spoilage benchmark and the early warning method during banana storage are effective.

Automated summary

The study established a unified spoilage benchmark for two batches of bananas by fusing data from different indicators and selecting feature wavelengths, enabling early warning of banana spoilage using Mahalanobis distance.