This project aims to develop artificial intelligence (AI)-basedalgorithms for automatic detection of pests in high water content plantsand fruits with x-RAY absorption (ABS), differential phase…

This $268,500 cooperative agreement from the Department of Agriculture Animal and Plant Health Inspection Service will support the development of artificial intelligence algorithms for the automatic detection of pests in high water content plants and fruits. Awarded to Stanford University on August 12, 2022 under the Plant and Animal Disease, Pest Control, and Animal Care program (CFDA #10.025), the project aims to modify an existing X-ray phase-contrast imaging system to enable continuous imaging of samples on a conveyor belt. Efforts will also focus on developing AI algorithms capable of detecting pests in plants and fruits with over 80% accuracy. Deliverables will include reports on deep learning performance and comparison with human experts. Findings are intended to benefit plant quarantine inspection and food safety efforts through AI-assisted detection of agricultural pests.

This project aims to develop artificial intelligence (AI)-basedalgorithms for automatic detection of pests in high water content plantsand fruits with x-RAY absorption (ABS), differential phase contrast(DPC), and dark-field (DF) images in combination. Activities willinclude modifying the vertically-orientated x-RAY phase-contrast imagingsystem we will finish at the end of phase i with an imaging systemcapable of continuous imaging using a conveyer belt to feed imagingsamples, updating the imaging system with the capability of a 12-15 cmlateral imaging field of view (FOV) and up to a spatial resolution of 20ΜM, and developing artificial intelligence algorithms for automaticdetection of pests in high-water content plants or fruits with anaccuracy of > 80%. Deliverables will be reports on the deep-learningautomatic infestation screening performance and comparison withmulti-reader studies. The reports should include details of the imagingprotocol, deep-learning algorithm development, key image featuresdetermined as infestation by the deep-learning algorithm, and accuracymetrics of the deep-learning algorithm and aphis professionals. Effortswill benefit plant quarantine inspection to detect insect pests inagricultural products with AI assistance. The findings and technologyfrom this research will also benefit food inspection of USDA andindustries.