How a KI revolutionices the world of forestry and insurance – the startup OCELL for smart analyses on aerial photographs
A forester goes into the forest and assesses the stand of his trees. He can rely on his training and experience. This is more important in 2020 than ever before, because the dry summers of 2018/2019 have particularly affected the native spruces. The loss of quality of this raw material is already having an impact on the international timber trade, as a result of which thousands of hectares of stands must be regularly monitored for damage and infestation by bark beetles. Normally this would require a lot of manpower, which is associated with high personnel costs, inconsistent quality and a lot of time.
THE DIGITIZED SOLUTION FROM OCELL
The Pain Point of the forestry industry, which becomes more acute at this point, is taken up by the StartupOCELL. The three founders offer an AI-based analysis of aerial photographs, but why did they choose the German forestry industry as a use case? “The forestry sector was simply a great starting point for us. We can approach people who have a problem and are desperate for solutions,” David Dohmen of OCELL told us. The startup has a large network of pilots who take aerial photographs for them, which are then analysed by the MachineLearning software developed by OCELL: The guys don’t know much about forests, but: “We in the forestry industry have little experience with Big Data and Artificial Intelligence. They help each other here. We were able to learn a great deal from our pilot customers and today we offer a product that is very well received and even recommended,” says David. In the forestry industry, aspects such as tree species distribution, tree heights and potential growth are important. This also brings up topics such as CO2 uptake: So if a levy comes, you can determine how much CO2 the forest is actually consuming. This is one of the topics of the future, together with early bark beetle detection, in order to be able to act in this area before the entire forest is gone: “The OCELL software has already proven itself in the field.
THE PROOF POINT OF THE STARTUP
“On the one hand, we classify the trees. We can tell exactly where each tree is located and from the heights we can estimate age and volume of wood.” David explains that OCELL is also able to analyse at what age the tree grows how fast. It is also possible to use the analysed data to make a statement about how many will grow next year. The AI developed by OCELL is also fully networked. The very good data situation can be combined with a very good scientific basis. “We also work with harvesters. These are the machines that cut out the trees. Their data is fed back directly into our models. This is all training again, and we are constantly improving the model with each new customer.”
AS ALWAYS: THE MAIN THING IS TO BE WELL INSURED
In the forestry industry the start-up was very well received, the next market they want to conquer is insurance: “This is a completely different industry. But here too, there are relevant findings to be extracted from aerial data and offered efficiently. The customer is interested in two aspects at this point: Firstly, the previous risk – for example, what is the risk for a building. If you have a normal roof or a flat roof, insurance for the latter is more expensive, which hides much potential damage and risk. According to David, this information is not available to the primary insurers or is very inadequate. OCELL’s approach is to pass on the building parameters such as flat roof, pitched roof, nearby trees, solar cells or even the material quality of the roof to the insurance companies via the property locations. The second aspect deals with the damage analysis afterwards. If, for example, three locations are hit by a storm, OCELL flies over it and takes the pictures with an aircraft that can also fly under the cloud cover – unlike the satellite – and records the data. This is also where OCELL’s great advantage lies: aerial photographs and analysis from a single source.
OCELL AND THE FACTORY1
Two of the three founders of OCELL have a date science background and come from a different startup. All three of them have been getting a taste of the famous “startup air” for quite some time, but have also gained experience in large corporations: “The motivation is to build up their own business, to be able to move more and faster. With our Data Science qualification we have the luxury of encountering many fascinating problems and being able to offer innovative solutions, which always drives us enormously”, says David.
The team became aware of WERK1 through the recommendation of Crashtest Security, another start-up that has grown up in WERK1. Both teams got in touch at an XPRENEURS alumni event. “They said they are sitting there now and love WERK1, so we simply applied directly. We are absolutely happy here too.”
We at WERK1 are looking forward to accompanying OCELL in its next development phase.