Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Hyperspectral Imaging for Identifying Foreign Objects on Pork Belly

Gabriela Ghimpeteanu, Hayat Rajani, Josep Quintana, Rafael Garcia

TL;DR

This work addresses the challenge of detecting small foreign objects on pork belly using hyperspectral imaging in the near-infrared range. It introduces a patch-based, lightweight Vision Transformer for per-pixel semantic segmentation of HSIs, complemented by flat-field correction and per-band normalization to mitigate temperature and lighting variations. The method achieves high detection accuracy with a loss-function designed to minimize false positives, aided by post-processing steps (erosion and spectral rules) to refine predictions, and demonstrates robustness to sensor temperature changes. The approach offers a scalable, real-time solution for automated quality control in meat processing, with potential to reduce manual inspection and enhance food safety.

Abstract

Ensuring food safety and quality is critical in the food processing industry, where the detection of contaminants remains a persistent challenge. This study presents an automated solution for detecting foreign objects on pork belly meat using hyperspectral imaging (HSI). A hyperspectral camera was used to capture data across various bands in the near-infrared (NIR) spectrum (900-1700 nm), enabling accurate identification of contaminants that are often undetectable through traditional visual inspection methods. The proposed solution combines pre-processing techniques with a segmentation approach based on a lightweight Vision Transformer (ViT) to distinguish contaminants from meat, fat, and conveyor belt materials. The adopted strategy demonstrates high detection accuracy and training efficiency, while also addressing key industrial challenges such as inherent noise, temperature variations, and spectral similarity between contaminants and pork belly. Experimental results validate the effectiveness of hyperspectral imaging in enhancing food safety, highlighting its potential for broad real-time applications in automated quality control processes.

Hyperspectral Imaging for Identifying Foreign Objects on Pork Belly

TL;DR

This work addresses the challenge of detecting small foreign objects on pork belly using hyperspectral imaging in the near-infrared range. It introduces a patch-based, lightweight Vision Transformer for per-pixel semantic segmentation of HSIs, complemented by flat-field correction and per-band normalization to mitigate temperature and lighting variations. The method achieves high detection accuracy with a loss-function designed to minimize false positives, aided by post-processing steps (erosion and spectral rules) to refine predictions, and demonstrates robustness to sensor temperature changes. The approach offers a scalable, real-time solution for automated quality control in meat processing, with potential to reduce manual inspection and enhance food safety.

Abstract

Ensuring food safety and quality is critical in the food processing industry, where the detection of contaminants remains a persistent challenge. This study presents an automated solution for detecting foreign objects on pork belly meat using hyperspectral imaging (HSI). A hyperspectral camera was used to capture data across various bands in the near-infrared (NIR) spectrum (900-1700 nm), enabling accurate identification of contaminants that are often undetectable through traditional visual inspection methods. The proposed solution combines pre-processing techniques with a segmentation approach based on a lightweight Vision Transformer (ViT) to distinguish contaminants from meat, fat, and conveyor belt materials. The adopted strategy demonstrates high detection accuracy and training efficiency, while also addressing key industrial challenges such as inherent noise, temperature variations, and spectral similarity between contaminants and pork belly. Experimental results validate the effectiveness of hyperspectral imaging in enhancing food safety, highlighting its potential for broad real-time applications in automated quality control processes.

Paper Structure

This paper contains 8 sections, 14 figures.

Table of Contents

  1. Introduction
  2. Image acquisition setup
  3. Computer vision software
  4. Data acquisition
  5. Hyperspectral image processing
  6. Machine learning model
  7. Results
  8. Conclusion and future work

Figures (14)

  • Figure 1: RGB images highlighting the similarities between a piece of PEHD and a blood stain on pork belly.
  • Figure 2: RGB images depicting the ten classes of contaminants used for this study: PA-PP (PA and PP were grouped into a single class due to the similarity in their spectral signatures), PU, metal, PEHD, Teflon, nitrile, wood, paper, cardboard, and white conveyor belt.
  • Figure 3: Examples of two hyperspectral images: one with fat side up (top row) and the other with meat side up (bottom row). Each image depicts seven of the 184 spectral bands, acquired at wavelengths 1012nm, 1113.5nm, 1218.5nm, 1323.5nm, 1428.5nm, 1533.5nm and 1638.5nm.
  • Figure 4: Examples of annotated data.
  • Figure 5: Hyperspectral mean intensity curve plots for our 13 test materials after flat-field correction, before (left) and after normalization (right).
  • ...and 9 more figures