Introduction: On-farm automated counting, identification, health assessment and weighing of pigs before
loading to livestock transports by deployment of computer vision and AI is requested by pig trading
companies, farmers and authorities. Around 34.9 million pigs are traded and transported annually between
EU member states and 0.5 million to or from non-EU countries. The reason is deficits of piglets in e.g.
Poland, Germany, Italy, Hungary, Rumania, and specialization in efficiently breeding and raising pigs by
countries like Denmark, Spain and Netherlands. Through the computer vision and AI based monitoring of
pigs prior the (un-)loading, the process becomes more efficient and flawless. The inspection by
veterinarians and drivers of the livestock transport unit can be based more on data than the current visual
and thus laborious inspection. Accordingly, the health and safety of the pigs being traded can (expectedly)
be improved. The pigs that are not in condition for transport can be selected at an earlier and more
appropriate time, avoiding the need for destruction of piglets.
Methodologies: Review of literature and state of the art hardware and AI system setup to fulfill stakeholder
specifications (derived from existing COMMECT data (COMMECT - Grant Agreement no. 101060881)).
Review of deep learning algorithms enabling the remote detection of tail bites, hernias, ear injuries and skin
lesions on pigs, as well as identification, counting and weighing of pigs. Studying optimal camera types and
their physical placement for full visibility of pigs prior loading in pig pens and during loading/unloading from
livestock trucks.
Output: A report presenting and discussing the road map for the hardware and AI engineering to obtain the
above-mentioned automation. An attempt will be done to collect a dataset of images of pigs in a pen and
manually annotate key health indicators such as lameness, ear injuries, tail bites, and hernias.
Impact: A road map for research and development towards harmonization and automation of the haulier
and veterinary inspection responsibilities in connection with livestock transport.
What I can offer: The above mentioned COMMECT project published a business model based on the overall
technology description, as well as conducted a stakeholder survey. The latter also resulted in a network of
all relevant stakeholders that will be involved in the validation of the UC output. The research group at AU
has years of research experience in computer vision and AI and have access to an experimental pig stable
facility at the AU Viborg campus.
What I need: Cooperation with experts on computer vision applied to animal science. This is to proceed to
the next step of applying for research and development funding based on the reporting from this UC and the
COMMECT project, and consortium formation.
Source: Annual AgData Use Case Summit 2026 booklet.
