A2-Inspect - Autonomous net inspection

Sorry for lack of information here, this page is still under development.

Sorry for lack of information here, this page is still under development.


Our autonomous net inspection enables a new level of control and security to fish pen nets by automatically detecting holes and weaknesses in the nets. Keep the fish safe and secure by controlling its environment

How to use (Read this)

The drone automatically inspects the net and reports back any findings, so if there is a hole or any damage in the net, you will get an alarm and report as soon as the drone has performed a net inspection. The interval of how often an inspection are performed are fully controlled and determed by you in the app, so the intervall really just depends on your needs. We would suggest starting with inspection every secont week, and after any operation where the net have been handled. But if you want a daily inspection this is all up to you. When not inspecting Probot would be cleaning or harvesting water-data.

How do the Net-Inspection work?

We use three main consepts, navigation, detection and reporting. This is the core that enables all our functions to work and give you as a fish farmer true value. The most important is the navigation concept, and for this we use our own patented system, a geenius consept where we calculate a distance and angle to a net (or it can be an other structure like a tank, acctually any confined space) and then we use this data combined with the inbord sensor data (depth, heading, roll, pitch, yaw) to position our drone relative to the structure we operate in. This patented navigation consept is what enable us to navigate freely in any confined space without any restrictions. This also makes our navigation system robust and takes away any need for surface inputs like USBL/INS.  For the detection we rely on data from multiple optical sensors that kick off our neural networks where machine vision detect whatever we train it to detect. This detection platform enable us to detect whatever we want, and also make our platform able to easily build new functions like pellet-detection, biomass, lice-counting, fish-healt, deadfish-counting and basically anything you need to detect. This is a first step to automate the sea stage of fish farming. As a bonus it also means you can finally start to get rid off stuff from your pens and use only one system for it all.

Summary of how it work:

  • Confined Space Coordinates: Adds accuracy to the y-axis. (patented metod)
  • Distance and angle calculation: More precise x-axis coordinate.
  • Depthsensor: Delivers more precision to the z-axis coordinate (mm level precision).
  • IMU sensor: Accelerometer thats adds even more precision to all axis.
  • Neural networks: Runs the machine vison for detection
  • Report and control code: Automatically sends the data from each mission with pictures
  • Server and cloud: Compiles the data from the drone and saves it for future BIG data value.
  • Webapp: Compiles the data from each mission and generates userfriendly reports.

Combining with the cleaning function?

So the Inspection and cleaning are two separate functions but we are working on integrating the inspection function to the cleaning function so everything can be perfomed in one operation. The problem are that when cleaning the net the cameras are so close to the nets that we get a bad overview, and then also a lesser inspection. It's inportant for us that every function we release are thrustworthy, fully tested, reliable and robust, so thats why we startet with inspection and cleaning as two separate functions. But for sure this is something we will merge together in a near future. We also have a lot of feedback that there are scenarios where you need to have a complete inspection and you dont want the ekstra time spent on cleaning. This can for an example be in the winter season, where algae on the nets dont grow as rappidly as in the summer(in norwegain conditions, this depends on where on earth you have your fish farm located) or after a storm.