OptoScale
An underwater AI camera used by Norwegian fish farms — weighs fish and watches for sea lice
The hardest question in fish farming is: how many fish are in the net pen, and how much do they weigh? In the past, this relied on sampling by scooping them out and weighing them — both harmful to the fish and inaccurate. OptoScale submerges cameras underwater, using imaging to directly estimate individual weight and biomass while also watching sea lice and health metrics.
Features and application scenarios
The product is called Bioskop 3.0, emphasizing Plug & Play automated deployment that integrates multiple data streams into one portal: weight and biomass, sea-lice count, health metrics, plus behavior analysis like swimming activity, respiration rate, and feeding response. All data can connect to the existing fish-farm management system via API, so you don't need to open another siloed system. Applicable species cover salmon, rainbow trout, smolt, and submerged farms. The company is headquartered in Trondheim, Norway.
Suited to commercial salmon/trout-farm operators and production managers. Taiwan's aquaculture is dominated by grouper, giant grouper, whiteleg shrimp, and tilapia — species different from OptoScale's model-training targets, so directly transplanting it won't be accurate — but the logic it demonstrates is worth learning: replacing "sampling estimation" with "continuous measurement" makes the accuracy of feed-conversion ratio and shipping scheduling a completely different level. For Taiwanese operators to use it, they'd first need to discuss local-species model calibration with the vendor.
Main features
- Underwater camera real-time measurement of individual weight and whole-farm biomass
- Automatic sea-lice counting and health-metric tracking
- Behavior analysis like swimming activity, respiration rate, and feeding response
- Bioskop 3.0 single portal integrating multiple data streams
- API connection to existing fish-farm management systems
Common uses
- Offshore net-pen biomass estimation and shipping scheduling
- Sea-lice monitoring and de-lice timing judgment
- Feed-amount optimization
- Long-term recording of fish-welfare metrics
Key Features
- Underwater camera real-time measurement of individual weight and whole-farm biomass
- Automatic sea-lice counting and health-metric tracking
- Behavior analysis like swimming activity, respiration rate, and feeding response
- Bioskop 3.0 single portal integrating multiple data streams
- API connection to existing fish-farm management systems
Pros
- Estimates weight without scooping fish, low stress on fish and high data frequency
- Sea-lice monitoring directly corresponds to Norwegian regulatory reporting needs
- Automated deployment, low operating barrier for on-site staff
Cons
- The model mainly targets salmon/trout; other species need separate validation
- It's a hardware + subscription model, with a non-trivial single-farm adoption cost
- No clear local agent and after-sales support in Taiwan yet
Use Cases
- Offshore net-pen biomass estimation and shipping scheduling
- Sea-lice monitoring and de-lice timing judgment
- Feed-amount optimization
- Long-term recording of fish-welfare metrics
Editor's Note
Aquaculture's digitization has long lagged agriculture, and OptoScale's approach of 'putting a scale in the water' is the most pragmatic one I've seen.
FAQ
Can it be used for Taiwan's grouper or whiteleg-shrimp farming?
Currently the model mainly targets salmon/trout, and directly applying it to grouper or shrimp has no accuracy guarantee. If you want to adopt it, be sure to confirm with the vendor whether it supports your species, or discuss local-data model fine-tuning — don't order based on the spec sheet alone.
Does murky water affect measurement?
Yes. Image-based measurement is inherently affected by visibility, and murky water lowers the recognition rate. The visibility conditions of Norway's cold-water seas are relatively friendly; Taiwan's nearshore and inland fish ponds have very different conditions, so this is the most important on-site test to do when evaluating.