In a span of three years, Pix4D has made its presence felt worldwide by coupling its solutions with UAV hardware manufacturers. The company’s CEO, Christoph Strecha, shares here his thoughts on the global UAV market and his vision for Pix4D.
Pix4D is a spin-off of a research institute in Switzerland. How did it all start?
We started from a university project more than 10 years ago. I was doing research on how to fully automate photogrammetric image processing workflows but with the same survey-grade accuracy. The existing workflow at that time was very manual, time consuming and specialised, which means only specialists could do it. I met a group working on computer vision and it was one of the leading groups looking at this technology from a computer vision point of view. So the idea was to combine computer vision techniques with photogrammetry. We developed software that is automatically able to process images into 3D models at survey-grade accuracy and, in 2011, we officially started Pix4D to sell our solution.
What are your focus areas of applications?
Right now our biggest focus is to provide solutions for UAV images, which we believe is a big market. UAV images are not of the best quality because a UAV needs to be light and safe, and the camera used is usually a consumer-grade camera. The focus of our software is to use the survey- and consumer-grade cameras with a high redundancy in order to have results that are comparable to HD cameras flying on manned aircraft. Since UAVs fly much lower, the image quality is not so good, but with the redundancy we can take images of a small area with the same accuracy as an HD camera on a manned aircraft. This is very interesting, especially for smaller areas like farming, because the cost of data acquisition, processing and time is substantially reduced. Other than UAVs, we’re also working in the general image reconstruction sector. We have solutions based on ground-based cameras that you can put on a vehicle. So we are not totally dependent on UAVs but this is where most of our business comes from right now.
What is your business model?
We sell software, both to UAV and camera manufacturers. We also sell to the end customers directly or through our global network of resellers. We don’t sell services. This is something that we’re also looking at but right now we’re making easy solutions that are usually coupled with hardware–that is, UAVs. We also have solutions that would use consumer cameras with fisheye lenses for instance, and we coupled these together to have an easy-to-use precise surveying tool. Most of our current revenue comes from coupling our software with UAVs. We also provide solutions for surveyors to survey small areas such as construction sites. We partner with UAV companies that focus on surveying or agriculture for instance. Then, together with them, we provide end-customer solutions.
Big companies like Trimble now have their own UAV and software divisions. Where do you see yourself in the market among big companies?
Trimble was the first of the big companies to acquire a UAV company–Gatewing, a Belgian company–which we were in contact with before and had a good business with. In fact, some of their users use our software to process data. So the big companies–if you’re talking about Trimble, Leica and Topcon— have realised that UAVs are important for mapping purposes; they have acquired hardware solutions for this and our software is also partially used for these UAVs. I think the acknowledgement by Trimble, Leica and Topcon of the importance of UAVs has given a push to the market which is still at a very early stage. This is a good sign for the overall market and we all need to work on making people see UAVs as a mapping tool just like they see laser scanners as a mapping tool. I think very soon every surveyor will own a UAV and, potentially, our software. So this is very good for us.
Do you think the market is big enough for the big players and SMEs to co-exist?
Of course. Trimble has a software division that focuses on software for traditional photogrammetry. This is very complex software, made for professionals, where you need to have a very good understanding of the software to get good results. Our focus is more on making it possible for the software to be used by anybody in the company, not just professionals. This is the advantage that Pix4D has and this is what contributes to its current success. I believe the three big companies will also work on this and, at some point, will be able to provide general solutions like us. We’re also now looking to diversify, to make specific products for particular problems. We have a big user base and we know what people want to do and what solutions they’re asking for. So we’re now looking at providing those solutions. I think we’re in a good place in the overall market structure. If you look at Leica and Topcon, they don’t have software solutions to process those images, so we are also collaborating with the big players to couple our software with their hardware. So that is good business for us as well.
Which verticals and geographies do you mainly cater to?
Our main clients are surveying companies, working for the mining, cadastre and natural resources management sectors. More and more agronomists and farmers are also using Pix4D, especially in USA. So this is a big market for us and is the area where we provide end-customer solutions. Our technology is also gaining importance for archaeology and cultural heritage.
We operate worldwide. Typically, regions where we have most of our success are the regions where it is easy to fly UAVs. Some countries already have a more liberal legislation than other countries. We have good business in Europe, Canada, South America, Australia, Japan, and China. We have a lot of business in these countries because it is easy to fly UAVs and we’re basically the main provider of UAV processing software there.
Do you organise any workshops or training sessions to update your users about your software and solutions?
We organise free Webinars on a monthly basis that cover specific topics of our technology. In addition, we also organise user workshops twice a year that cover not only the more advanced use of our software but also topics such as sensors, type of cameras, UAVs and their use for specific goals. I think there is quite a demand on this.
What are your diversification plans?
Our software’s main function is turning images into 3D models in an automated way and focusing on survey-grade accuracy. We want to have a surveying product which is highly automated. That means that we can basically produce 3D models from any camera as soon as they have sufficient overlap. It could be consumer camera put on any UAVs with GPS tags. We have now launched a new product that can process 180 degrees fisheye cameras, which are really interesting for indoors where you don’t have any GPS and you’re not easily able to geotag your images. But with this product, you can use ground control points to georeference the models that come out of the computation. Normal Canon or Sony cameras with fisheye lenses can be used. They’re also very interesting in the urban environment. When you walk through cities, you don’t have a wide possibility to place your camera where you want because you’re restricted by the space. The fisheye lenses will give you a full overview and we can get highly accurate 3D models from them. So this is one thing where I believe we will have more business in the future – computing orthomosaics of facades and making building models inside and outside, etc.
Since we have strong knowledge in image processing and this is our main competence, whenever we’re looking to diversify, it will be in this direction. Now that we already cover lightweight UAV imagery as well as ground-based imagery without geotags, we are looking into new solutions that are more based on time modelling. Pix4D–the name itself shows what we’re doing partially today but will also do in the future; we’ll be analysing 3D models over time. For instance, in building construction sites and mining industry, the time factor is important. So there will be more analysis in that direction.
Is there any plan to buy a UAV company or manufacture your own UAV?
We have been thinking about this but what we are really looking for is to keep our focus on the software and have good partners that are able to produce the hardware. And that would be for any market. We are partnering a lot with companies that experienced in hardware. We are partnering with them to offer the end-customers fully integrated solutions combining both hardware and software.
Hardware is a business that is local; hardware can break so you need to have spare parts available locally. People want to talk to a local person if something does not work. Today, there are several UAV companies, some bigger than others. But I think the local aspect of a UAV company is quite important.
Where do you see Pix4D five years from now?
We have seen big changes in two years. There are 40 people working in Pix4D now. Two years ago we were only a team of five people. We’ve opened an office in Shanghai (China) and San Francisco (USA) this Fall, mainly for business development but also for dedicated solutions for farming, construction, etc. So I think even in two years Pix4D will be a lot different than it is today. In five years I hope we become a leading software provider for photogrammetry image processing. That’s our goal. To a certain extent we are already there but the market is growing and I hope Pix4D will have a big place in that market.