§Just one of our regular days at work. Flying blimp over the Lake of Neuchatel and Jura mountains. A few weeks ago, on a Friday night the most adventurous part of the Dotphoton team set off to Neuchâtel, beautiful town in Switzerland to wake up at 4:30 a.m. the following morning to run a special project.
Over the past couple of years we have been working in partnership with the European Space Agency (ESA). The project started in 2020 to adapt our image compression technology to space, followed by a contract renewal in 2021 to extend the work. Eventually, the goal is to have our flagship raw image compression product, Jetraw Core FPGA on satellites. What makes a “space grade” technology is in large part how reliable and well tested it is. The maturity of the product is measured in terms of “Technology Readiness Level” (TRL). They range from TRL 1 to TRL 9, where TRL 1-2 cover the concept and TRL 9 is where the technology gets approval after a successful completion of a flight mission in space.
Technology Readiness Level (TRL) scale levels applied to ESA's Technology Programmes
Dotphoton has previously completed TRL 1-4, having validated Jetraw’s high performance for speed and quality. We are now at TRL 5-6 and must show that the technology works in an environment as close as possible to the foreseen missions. This is as far as one can go without targeting a specific satellite.
For Jetraw Core, the environment includes the electronic interfaces, the chip on which it runs, the properties of the sensors and the image content which is acquired. We needed images of rural, agricultural, urban, suburban and water to test, and beautiful Swiss Neuchâtel region offers all of these subjects. Furthermore, we needed to have the best possible images, as images can in general be artificially degraded to emulate lower resolution satellites, but not improved.
Dotphoton’s R&D team built a self-contained, battery powered optical system that continuously acquired images with no user intervention. To maximise the sharpness of the images the system had to be mechanically and thermally very stable.
The system was installed on an electric blimp and included two cameras made by Baumer with Sony IMX530 CMOS solid-state image sensor; one “panchromatic” (monochrome) and one “Bayer” (colour). Both to deliver raw data from the sensor.
The goal for this ‘mission’ was to take reference images which would later form a baseline for generating synthetic data. The latter implies artificially creating images as if they were made on satellite, which could then be used to test Jetraw compression.
First, the system with two cameras, batteries, embedded computer weighs 10kg. It would require a big industrial drone. Not only they’re expensive, but they’re dangerous if something goes wrong and they fall down.
Second, drones are remote controlled, and electromagnetic interference can result in the loss of the drone, with complex electronic systems such as the one we built, even with best practices there is always the risk of interference between the system itself and the drone remote control, especially if you have no “ground” connection.
"What about those drones with cameras installed in them, couldn’t you use those?" - you may ask. This is what we had done in the first part of the project, however the sensors on those drones are consumer grade, and do not offer the same features as the ones used in satellites, such as global shutter. Also, the optical stabilisation system on those drones, although good for consumer purposes was not sufficiently stable for our tests, which need a stability of better than a tenth of a pixel. Also, they don’t come in monochrome. This is why we decided to build a perfect optical system ourselves and make it fly somehow.
A plane could work well, but is quite fast and has high levels of vibration. We needed to capture very high quality images, meaning long exposure times, which on a plane would have resulted in motion blur.
The hot air airship we used provided a perfect solution: it has high load capacity, minimises vibrations, especially the one we flew, which is a prototype of an electrical powered airship, and therefore has even less vibrations and no exhaust gasses to damage the images.
Generally, it takes pilots about 15-20 hours to learn how to fly a plane, while it requires 50-100 hours for them to go on a solo trip on an airship.
Thermal airships can only fly few times a year due to environmental requirements: no fog, low wind, zero probability of rain, and with a sufficient time frame, as it takes a long time to set it up. The team started to put blimp together at 5:30am and could only take off at 7am. After two flights, it was another 1,5h to deflate it and pack it back.
It was, the day was gorgeous, we did two 1-hour flights, and the system took about 800 images in colour and b&w. Now we’re working on generating the synthetic data and running it through Jetraw, thus moving closer to TLR-6.
Overall, we’re so excited and grateful to apply our team’s out of the box thinking and multidisciplinary skills daily. It’s such a privilege to be working on different innovative technologies, which are aimed at enhancing science research, daily operations and the lives of people globally.
We could not have done this without a few people. Mr Fabien Droz, who engineered and piloted this innovative electrical powered hot air airship taking Dotphoton system on board of this wonderful vehicle. We also thank Yosef Akhtman, who introduced us to Fabien. And the two brave Dotphoton team members Nadja and Phillip who volunteered as 2nd pilots for the two flights, guarding data acquisition equipment and taking beautiful shots from the sky featured in our short video you can watch on our YouTube. Lastly, everyone who helped set up and unfold the blimp, including Dotphoton's office dog named Mango.
