AI Fish Size Estimation Camera

Deep learning image recognition for fish detection

"AI Fish Size Estimation Camera" is a system that captures fish using a stereo camera, detects fish bodies through deep learning image recognition technology, and calculates the length and height of the fish based on their position information.

This system allows for the measurement of fish size and weight without removing them from the pond. Data on fish body size and weight, including 'average', 'maximum', 'minimum', and 'standard deviation', can be viewed in a list format on a computer or smartphone. In addition to listing the data, the system enables the accumulation of video data and the visualization of measurements over time through graphs, making it possible to monitor growth processes in real-time and analyze accumulated data.

Furthermore, since learning models are created for each fish species, the system can measure various fish species without being limited to specific ones.

Patent granted　"Fish Size Calculation Device"

"AI Fish Size Estimation Camera" combines deep learning and stereo camera surveying technology to estimate fish size without physical contact. This device uses an underwater stereo camera with two cameras positioned side by side to capture three-dimensional images of swimming fish. Fish that appear sideways in the images are detected using deep learning object detection technology. The size of the detected fish is calculated using the position information of the fish and the principles of the stereo camera.

At the time of patent acquisition, the measurements included 'body length' from the tip of the upper jaw to the base of the tail fin and 'body height' from the base of the dorsal fin to the base of the pelvic fin. Currently, it is also possible to measure fork length, total length, and weight.

Use Cases

• 

  Land-Based Aquaculture

  Atlantic salmon

  Denmark       Danish Salmon

• 

  Offshore Aquaculture

  Sea bream

  Minamiuwa, Ehime Yasutaka Suisan

• 

  Offshore Aquaculture

  Sea bream, yellowtail,
  amberjack, striped jack

  Nagashima, Kagoshima

Developing Aquaculture, Ensuring Food Safety, and Stability

By utilizing this technology, it becomes possible to measure fish without physical contact, thereby avoiding stress on the fish and reducing the risk of mortality and disease, which can lead to higher added value. Optimizing feed amounts can reduce costs, streamline production activities, and improve measurement accuracy and productivity without relying on the skills of individual workers, contributing to the resolution of workforce shortages.

At i-ocean, we are committed to achieving the development of aquaculture, food safety, and stable supply. We continuously provide new value to society through our camera and analysis system for measuring fish size, 'AI Fish Size Measurement Camera', our 24-hour water quality management system, 'IoT Water Quality Sensor', and our 'Aquaculture DX Support', which promotes automation and labor-saving in the fisheries industry, thereby contributing to a sustainable society.