Development of a Hyperspectral Imaging System for the Early Detection of Apple Rottenness Caused by Penicillium

Abstract

AbstractAutomatic detection of early rottenness caused by Penicillium is still a challenge due to the high similarity between the rotten region and sound tissue, and the large variations in surface background color of apples. A hyperspectral reflectance imaging system and a systematic approach were developed to detect the early rottenness caused by Penicillium on Fuji apples. The principal component analysis of the hyperspectral reflectance images in the full region (400–1,000 nm) of spectra was used to develop hyperspectral image analytical algorithm and select the efficient wavelengths. Independent component analysis (ICA) based on the selected efficient wavelengths (810 and 970 nm, respectively) and simple thresholding methods were used to segment and classify the rottenness from the sound tissue. Finally, ICA‐based image processing and classification algorithm was developed based on the images at 810 and 970 nm to realize the fast inspection of early rottenness. The detection result with 97% overall accuracy indicated that the proposed system and algorithm were efficient and suitable for the detection of the early rottenness caused by Penicillium on Fuji apples. This study was the first attempt to use hyperspectral imaging system to detect the early rottenness caused by Penicillium on Fuji apples and would lay a foundation for the development of multispectral imaging system for fast inspection of rottenness. The limitation of our research is the limited inspection view of the samples. Future work will be focused on simultaneous whole surface inspection using multispectral imaging system.Practical ApplicationsThe algorithm and results in our paper can provide a foundational basis to develop a real‐time inspection system to detect rottenness on apples in the postharvest processing line, or develop a monitoring system to monitor the quality changes in storage shelf.