Stereo Camera System Estimates Fish Weight
Researchers at Sultan Qaboos University developed StereoGeoBox, an underwater system that estimates fish length, width and weight from paired camera images.
Updated on Nov. 7, 2026 in Aquariums
Researchers at Sultan Qaboos University developed StereoGeoBox, a fish weight estimation system that uses two synchronized cameras and YOLO11 artificial intelligence. The study was published in Computers and Electronics in Agriculture on October 23, 2026.
Why it matters
Estimating fish weight without handling the animals could help aquaculture operators monitor growth more efficiently. The system's processing speed and low latency point to real-time use in underwater environments.
In a controlled test, the system estimated a 15.30 centimeter tilapia at an average of 15.25 centimeters while running at 25.40 frames per second with 28.2 milliseconds of latency on an NVIDIA Jetson Thor platform. The system's GPU memory use was 17.7 megabytes, and its accuracy beyond tilapia remains untested.
The players
Sultan Qaboos University
A public research university in Oman, founded in 1986, that conducts research across engineering and agricultural sciences.
Abdullah Al Salem University
A public university in Kuwait that collaborated with Sultan Qaboos University on the research.
The details
The system calculates depth from the geometric relationship between fish detected in paired images, then feeds estimated length and width into an XGBoost machine-learning model to predict weight. Researchers evaluated it using tilapia monitored over 15 months at four growth stages, with manual measurements as reference values.
Timeline
October 23, 2026: Study published in Computers and Electronics in Agriculture.
Over a 15-month period: Tilapia monitored at four growth stages during the study.
Under the Hood
The system applies YOLO11, a real-time object detection model, to underwater stereo imagery to estimate fish dimensions without generating a full depth map of the scene. This approach extends detection-based vision methods into aquaculture measurement, where prior systems often required more computationally intensive depth reconstruction.
Aquarium keepers and fish farm operators could eventually use camera-based systems like this to track fish size and weight without netting or handling the animals. The prototype is still experimental, and its practical availability for hobbyist or commercial tanks has not been established.
The takeaway
Camera-based monitoring could reduce stress on fish by replacing manual weighing. Its usefulness beyond tilapia and controlled test conditions will depend on further validation.
Further reading
Explore more on Aquariums for context on underwater fish care and monitoring.
Source note: This article includes information reported by Ocean News & Technology.
