Who would’ve thought AI and Comp. vision can help save fishes by Cannoning them?

03 May Who would’ve thought AI and Comp. vision can help save fishes by Cannoning them?

Posted at 16:27h in Blog by Rinkesh D

Banner image generated using ChatGPT, OpenAI May 1, 2025

Dams have a huge environmental impact. Yet they are also important for flood control, irrigation, and Hydropower generation. Hydropower is the most reliable source of renewable energy. In fact, as per ‘Our world in Data’ as of 2023 it constitutes ~14% of Global Electricity and ~8% of India’s electricity is generated by Hydropower [1].

Fishes, especially migrating fishes like Salmon are among the most affected by large dams [2]. As per Northwest Power and Conservation Council “More than 40 percent of the spawning and rearing habitat once available to salmon and steelhead in the Columbia River Basin is permanently blocked by dams.” A similar impact is seen in India too. And the issue is not purely ecological too. As per a report by Asia Network on Dams, Rivers and People, “mighty River like Brahmaputra supports more than 2 million fisher folk, smaller rivers like Wainaganga in Maharashtra support more than 4 lakh fishermen. The fisher folk in India can be termed as the poorest of the poor.” [3]

Clearly preserving the ability of fishes to migrate has huge ecological and socio-economic benefits. In countries like USA and Europe, this is mandated by law and regulations, some as old as 1890s [4]. In some cases this is done by dam removals, but same is not viable in every case and other solutions are needed.

A classic solution for this is a fishway or fish ladder, which help fishes when they need to move upstream via a “ladder contains a series of ascending pools that are reached by swimming against a stream of water. Fish leap through the cascade of rushing water, rest in a pool, and then repeat the process until they are out of the ladder” [5]. Fish ladder mostly works for anadromous fish such as salmon, which need to travel upstream as adults. However, they tend to have an efficiency of 75% for Salmons but 25% of non-salmon fishes [6]. New approach is needed.

Enter Fish Cannons, the viral videos of which have been popular on the net [7]. Do check out these videos, they are fun! [8]

Fish cannons (Whooshh Fish Transport System) helps launch fish via a flexible tube and with air pressure in a pneumatic tube, over the dam, at times over pretty long distances at speeds of 20 miles/hour. This system is created by Whooshh Innovations, a startup [9]. This is very effective, as only 3% of fish show some injury. But for the system to work, the fish must be of the right size as smaller fish, may not block enough air to generate the necessary pressure to be launched far enough [10]. This is the reason in early version of this system, fish tended to be loaded manually [11]. This of course is very costly and not scalable.

Hence, as addition, Whooshh Innovations, has created Whooshh Ellips Scanning Sorting system, which uses Computer vision and AI to sort the fish by size and even by species of fishes. Basically, when the fish enter the system, an image is taken of the fish. Then Machine learning algorithms are used to identify and sort the fish. To quote a detail explanation from the company “six cameras (three using near-infrared and three using visible light) capture 18 images of each fish from different angles. The images are then processed by AI algorithms that recognize the fish species, size, and other features, such as the presence or absence of an adipose fin.” Then these fishes are directed to the right size of tubes from where it can launch with necessary pressure. Some versions of the technology can also identify invasive species of fish, which cause harm to the environment and remove them. Overall, while not perfect the technology works (3 out of 4 times) and its further development will help scale the fish cannon to viable method to help protect fish and culture and economic wellbeing of people dependent on them [12].

While any and all development by humans will have a negative environmental cost, it is important that we mitigate these as far as possible. As the example of Salmon cannon shows, technological innovation is a key aspect competent of this effort. With proper support, cutting edge tech innovations like AI/ML and Computer vision can help us mitigate the negative impacts of development on the environment and society. It is essential that such innovations are promoted and supported by society.

Footnotes:

¹ https://ourworldindata.org/grapher/share-electricity-hydro?tab=chart

² https://www.nwcouncil.org/reports/columbia-river-history/damsimpacts/

³ https://sandrp.in/wp-content/uploads/2018/03/impacts_of_dams_on_riverine_fisheries_in_india_parineetadandekar_sept2012.pdf

⁴ https://www.fao.org/4/y2785e/y2785e03a.htm# ; https://nyuelj.org/2017/10/ferc-fish-pacific-northwest-dams/?; https://www.youtube.com/watch?v=MonfznEl1hk

⁵ https://oceanservice.noaa.gov/facts/fish-ladder.html ; https://www.youtube.com/watch?v=MonfznEl1hk

⁶ https: //oceanservice.noaa.gov/facts/fish-ladder.html; https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1467-2979.2011.00445.x

⁷ https://www.youtube.com/watch?v=EQ7pubFMb_0; https://www.youtube.com/watch?v=MonfznEl1hk; https://www.vox.com/2019/8/12/20801930/fish-tube-meme-whooshh-wfst-dams-environmental-impact

⁸ https://www.youtube.com/watch?v=2z3ZyGlqUkA; https://www.youtube.com/shorts/oYRDvpkLPa0

⁹ https://www.youtube.com/shorts/oYRDvpkLPa0, https://www.youtube.com/watch?v=EQ7pubFMb_0, https://www.vox.com/2019/8/12/20801930/fish-tube-meme-whooshh-wfst-dams-environmental-impact

¹0 https://www.youtube.com/watch?v=EQ7pubFMb_0

¹1 https://www.vox.com/2019/8/12/20801930/fish-tube-meme-whooshh-wfst-dams-environmental-impact;