This project will remove the number one bottleneck in the large-scale poultry slaughtering industry by the development of an advanced machine vision system able to replace the manual vet inspection in poultry slaughterhouses. The goal of the project is to develop a system ready for implementation at all Danish slaughterhouses and to export it world-wide. Since veterinary control is linked directly to food safety, there are strict requirements for a high detection rate and a low number of false positives. This coupled together with defects and diseases that covers all challenges from hard-to-detect fine-grained details to non-trivial holistic inspection of the whole bird, makes for a challenging computer vision task. This will push the boundaries of what is possible with today’s deep learning models for computer vision and has from a technical point to be solved involving spectral imaging. The developed vision system that involves public health, food safety and quality, will be validated in relation to manual internal and external inspection at selected slaughterhouses. It will be monitored by the Danish Veterinary and Food Administration's meat control department for the purpose of assessment as it will be presented for the veterinarian regulatory body in EU.