Image credit: Asimetrix
Hog farmers monitor pigs’ weight to track their health and growth. But weighing pigs can be a Catch-22. Weight determines when “finishers” are ready for market, but the act of weighing pigs can cause them stress, which can lead to weight loss, poor meat quality, and even death.
Because the risks are high, pig producers don’t typically weigh their pigs every day. In addition to the health concerns, it requires a lot of machinery, human interaction, and time. IoT is changing the game by allowing farmers to watch their weight, pig style.
High Above the Hog
Asimetrix has created PigVision to monitor pigs as they move freely in the barn or pen. Farmers install a PigVision camera in each pen, above the area where the pigs move around. The camera uses artificial intelligence and computer vision to take pictures of the pigs below and analyze the data.
The pigs don’t need a sensor or tag—all the intelligence is in the PigVision camera. “We have a processor that knows when the pig is under the camera. The camera knows when to take a proper picture of the pig,” says Juan Esteban Betancur Ochoa, lead technology developer and deep learning specialist at Asimetrix. “Every camera has Wi-Fi capabilities so it can connect through the Wi-Fi already on site or [farmers] can put in a mobile router if needed.”
The images are sent to the cloud, where proprietary neural networks and machine learning algorithms process the information and estimate the average weight of pigs in a given enclosure. The PigVision system doesn’t track the weight of individual pigs but rather provides an average daily weight gain per pen.
“The accuracy of the algorithm is between three and five percent—that’s very good,” says Juan Pedro Fernández Sanin, CEO of Asimetrix. The error rate varies depending on the genetic line of the pigs, he adds, but typically falls within the weight range for finishers.
On His Farm He Had a Pig and E-I-E-IoT
Asimetrix also offers sensors that track a range of environmental variables, such as temperature, humidity, ammonia levels, and water consumption, Sanin says. The sensor data is sent to the camera through a proprietary gateway, then on to the Asimetrix portal.
Farmers monitor pig growth through the Asimetrix dashboard, which provides real-time displays of the average daily weight of the pigs and any environmental information being tracked. Alerts can be sent via phone or email if specific thresholds are met.
Image credit Asimetrix
The dashboard also allows farmers to predict when finishers will be ready for market and operational changes that can improve efficiency on the farm. “We can detect anomalies in growth. We can detect when a pen is growing too fast or too slow,” Sanin says. Those anomalies could indicate health issues or illness within a pen or potential feeding issues—not enough or too much feed for certain pens.
A similar system, called Piggy Check, comes from Meier-Brakenberg. The Piggy Check application uses a 3D camera and AI-based software to determine the weight of pigs while they move about their enclosures. The software runs on a smartphone or tablet and combines the digital images or videos of pigs to create “depth images” that are used to estimate weight.
The weight is displayed on the phone right after the photos are taken, and the data collected is stored in the cloud for later analysis. When finishers are taken to market, the actual body data is compared to the Piggy Check data, so farmers can use precise statistics to make adjustments based on genetics and feeding regimes.
Knowing when finishers are ready for market can improve efficiency on the farm. For example, if the finishers are removed in a more timely fashion, the farm can operate more economically. They save money on feed because the largest pigs eat the most and they don’t overfeed pigs beyond their ideal weight. The benefit of computer vision translates directly to increased profits, allowing farmers to bring home the bacon.