Ultrasonic tomography for detecting foreign objects in refrigerated milk cartons

Abstract

This study presents an investigation of the detection of foreign objects in milk cartons using an ultrasonic tomography measurement system. Ultrasonic parameters such as acoustic impedance and wave amplitude were modelled and investigated using foreign materials of different densities such as metal, glass, wood and plastic. Linear back‐projection (LBP) and convolution back‐projection (CBP) algorithms with Hanning, Hamming and Ram‐Lak filters were implemented to reconstruct two‐dimensional cross‐sectional images of the milk and the foreign objects. The image quality assessment was conducted using the peak signal‐to‐noise ratio (PSNR) and structural similarity index (SSIM). The results revealed that the CBP‐Ram‐Lak filters managed to obtain the clearest and the sharpest reconstructed images as they achieved the best values of PSNR and SSIM (15.4537 and 0.8683, respectively) for a single glass material. It is envisaged that the ultrasonic tomography method can be very useful for monitoring the quality and safety of milk products, especially for detecting physical contaminants.