Unsupervised Foreign Object Detection Based on Dual-Energy Absorptiometry in the Food Industry

X-ray imaging is a widely used technique for non-destructive inspection of agricultural food products. One application the autonomous, in-line detection foreign objects in samples. Examples such inclusions are bone fragments meat products, plastic and metal debris fish, fruit infestations. This article presents processing methodology unsupervised object based on dual-energy absorptiometry (DEXA). A novel thickness correction model introduced as pre-processing DEXA data. The aim to homogenize regions image that belong product enhance contrast where present. In this way, segmentation more robust noise lack contrast. proposed was applied dataset 488 samples products acquired from conveyor belt. Approximately 60% contain different types sizes, while rest void objects. results show without correctly identified 97% cases overall accuracy reaches 95%.