The meat industry requires intelligent systems for the non-invasive, real-time detection of bone fragments.

The aim of a recent study, carried out by a group of international researchers (Collazos-Escobar et al., 2025), was to assess the feasibility of using ultrasound imaging and multivariate image analysis to detect such fragments in boneless, skinless chicken breast fillets.

For the experiment, small bone fragments of various sizes were introduced into the product, and images were acquired by scanning the surface using contact sensors operating in transmission mode. Various statistical techniques were then employed to classify the samples into those containing fragments and those free of fragments.

The results demonstrate that the presence of bone fragments in the product causes alterations in the parameters of ultrasonic energy magnitude (average decrease from 81.6% to 52.6%, depending on the size of the fragments) and distribution (average decrease from 97.9% to 70.6%, depending on the size of the fragments). It was also observed that the best sample classification performance is achieved using the RSS (Sum of Squared Residuals) statistical method.

In summary, the results obtained so far confirm the potential of combining ultrasound imaging with multivariate image analysis as a reliable and rapid method for detecting bone fragments in chicken breast.

Bibliography: G.A. Collazos-Escobar et al., Food Research International, 206, 2025, 116047