CyanoVision: Developing a low-cost, open-source early-warning system for cyanobacteria
Cyanobacteria are algae-like bacteria which inhabit freshwater, coastal and marine waters all over the world. These bacteria can quickly reproduce and form ‘blooms’ which are commonly referred to blue-green algae in water. Favourable reproduction conditions are hot climate with abundant sunlight, stagnant or slow-flowing water, and high nitrogen and phosphorus pollution levels (e.g. from sewage and agricultural fertilisers). These blooms can form toxic “carpets” on the water surface which can cause several environmental problems: they can disrupt drinking water supplies, negatively affect recreational activities and water-dependent industries, and pose a risk to livestock, wildlife and human health and life.
Our team of 12 students delivered a low-cost, open-source solution for early detection of cyanobacteria in water, before it forms into harmful blooms. Their CyanoVision technology takes samples in-situ and analyses them in real-time. This information can then be used to notify and warn organisations, such as environmental agencies, about the presence of cyanobacteria in water and indicate the levels of concentration of the bacteria.
This technology has potential for significant time and cost saving benefits through different practical applications. For example, with early detection, environmental agencies can advise on appropriate treatment and disposal of stormwater, agricultural, industrial and sewage effluent; plant or maintain riparian vegetation to sequester nutrients that cyanobacteria feed on; and/or manipulate water flow to prevent the build-up of blue-green algae.
The technology could also be used by individuals to perform the tests themselves and check whether the waters are safe for recreational activities when it would otherwise not be as visibly obvious.
How does the technology work?
The students have created a novel microscopy system based on the OpenFlexure Microscope for the detection of cyanobacteria and algae in fresh water.
Their machine learning algorithm, based on objected detection, has been trained to identify cyanobacteria species of different morphologies (e.g. unicellular, filamentous and rod-shaped) as well as the common algae in bright-field images.
They have created a user-friendly UI that enables the use of our software without requiring any coding knowledge.
More information can be found here: