By: James Mason, senior vice president of space systems at Planet
It’s been about six weeks since we launched Flock 4p, our latest batch of satellites, into space on an Indian PSLV rocket — bringing 12 of the latest-and-greatest SuperDoves to join our one-of-a-kind Planetscope global monitoring constellation.
The first imagery from Flock 4p was produced within a week of launch, like this stunning example collected on December 3rd, 2019:
To get to this point we’ve taken Planet’s unique iterative agile aerospace approach, launching technology demonstration and product pilot SuperDoves over the last year to iron out kinks in the design, prepare the imagery production pipeline, optimize operations and ultimately validate the product. The recent PSLV launch therefore marks a major milestone for Planet, as we start to transition the SuperDoves from this development phase into production.
At a glance the SuperDoves may look like classic Doves, but significant changes have been made under the hood. They can bring down about four times more data than the Doves, an increase that has allowed us to widen the swath (or field of view) by 30 percent, double the number of spectral bands (or colors) to eight, and improve the radiometry (how accurately we can measure the brightness of the light reflected back from the Earth in each pixel). We expect that this will significantly improve the precision and consistency of analysis or science fueled by this data.
In addition, these spectral bands have been chosen specifically to allow for interoperability with the Planetscope archive, the Landsat and Copernicus Earth observation programs, and the RapidEye constellation, the latter of which the SuperDoves will ultimately replace.
All that said, you might be wondering….
What happens between rocket launch and product launch to prepare satellite imagery for customers?
Today, we’re providing a behind-the-scenes look at the work that comes in the weeks and months following a launch to get things ready for market.
First Contact & Health Check
Immediately after launch, the first thing we need to do is make first contact — to establish communication between our mission control team and the satellite. This allows us to confirm that the satellite deployed from the rocket powered on and was delivered into the right orbit.
We do this by scheduling a first contact task in our automated mission control system, which was developed in-house over many years. This task waits for the first pass of the satellites over one of our many ground stations, and as the recently released group of satellites passes overhead, it tracks the expected orbit of the satellites and sends a “ping” message to each satellite over our encrypted radio links.
The satellites, on hearing this, respond with a cheerful “pong!” which is enough to tell us they’re working and in the right place. This back and forth ping-pong using our UHF radio (see the open source version here) can also be used to time how far away the satellites are, allowing us to estimate their initial orbits.
To date we’ve maintained a 100 percent success rate of contacting every satellite — all 363 of them — that we’ve deployed in space, so it is always a nail-biting time as we wait to hear back from each new sat. Typically we’ll hear back from every satellite on the first couple of passes and transition into health checks. This is where we assess the basic health of every satellite, asking:
- How charged is the battery?
- What are the spin rates?
- Are the temperatures what we expect?
- Are the solar panels producing some energy?
- Have there been any unexpected reboots?
If these automatic checks are all successfully completed, as they were on the Flock 4p launch, the satellites launch into their pre-planned commissioning procedures.
Detumbling, deploying solar panels and performing software updates
Now that the satellites are in the automated commissioning flow we’ve refined through many large launches, the next thing we need to do is “detumble” them and get them oriented properly.
From the deployment off of the rocket, the satellites are usually tumbling slowly, but we need the satellites to point very precisely to charge the batteries with our solar panels, communicate with the ground stations and capture imagery.
We do this initial detumbling using magnetorquers. While they sound like they could be Power Ranger villains, they’re actually electromagnets typically used to stabilize spacecraft. Magnetorquers exude magnetic fields that react to the Earth’s own magnetic field, creating a counterforce or torque. We use this counterforce to slow down the spinning of the satellites over a few hours. Once they’re slow enough, we can switch over to using reaction or “momentum” wheels that rotate the satellites by spinning a motorized wheel very quickly in the opposite direction.
As soon as the satellites are fully stabilized, we deploy their solar panels and orient these towards the sun, greatly increasing how much power they generate, preparing them for the power hungry activities like capturing and transmitting imagery.
At this point in the checkup, three months may have passed since these satellites were shipped for launch, so at this step we usually perform software updates over our encrypted radio links, upgrading everything from the flight software to the camera compression software with the latest and greatest.
Despite our satellites being smaller and cheaper than traditional, large remote sensing satellites, our goal is to create an imagery of equivalent quality, so how we calibrate our satellites is critical.
First, we need to calibrate all of the sensors onboard the satellite that allow us to point the satellites precisely, like the magnetometers (which measure the Earth’s magnetic field), rate gyros (which measure spin rate of the satellite), sun trackers (which measure direction to the sun), star trackers (which measure orientation based on imaging stars), and even the main imaging telescope.
We image with the main telescope at the same time as these other sensors to very precisely measure the angles between all of these sensors in case they’ve shifted minutely due to the launch or changing temperatures. Over the course of launching 300+ satellites, we’ve built automated procedures to collect and calibrate this sensor data. More details on this process can be found here.
Second, we need to compare the readings of each satellite and check their accuracy when collecting data. An image of the same location will vary based on the unique properties of each satellite’s sensor and filters, what the atmospheric conditions over the target were, how long the telescope has been in space, the temperature of the sensor and many more factors. Being able to correct for these sources of measurement error is crucial if you want to measure a physical quantity on the Earth when imaging from space (e.g. using our analytic surface reflectance product).
Finally, once initially calibrated, each satellite’s imaging performance is tracked over time and autonomously re-calibrated on a monthly basis to ensure any drifts or degradation is compensated for.
First light and the path to production
Although many images are taken for internal use during commissioning, it is only after calibrations and software updates are completed that we are ready to produce imagery products. The focus here shifts from the satellites to the actual imagery our customers will consume, so this stage includes quality assurance checks on the imagery and associated metadata, end-to-end testing of the imagery production pipeline (all of the software that gets an image from the ground station into our customer-facing platform) and collecting real customer feedback via an early access program (EAP).
The Flock 4p SuperDoves launched last year will finish this final stage soon, and their data is already in the hands of a subset of our Early Access Program customers prior to a wider release to all Planetscope customers. Over the course of 2020, we will continue to launch SuperDoves and will gradually transition our 3 and 4-band Dove users and 5-band RapidEye users over to interoperable products created from SuperDove’s 8-bands, as well as upgrading our global monitoring dataset to include all 8 of these spectral bands.
Every new SuperDove we launch will walk through the process outlined above, going through commissioning and calibration, before getting down to business to make their contribution to our mission to image the whole world every day and make data accessible and actionable.
Every day we are building upon a long term global record of deep stack multispectral time-series data, and we’re incredibly excited to see the insights our customers and partners will glean from the SuperDove data going forward. Stay tuned for the official product launch later this year!
If you’d like to learn more about our imagery products or to request sample data from the SuperDoves, please contact our sales team at email@example.com.
Also, our Planet team is growing and we’re always looking for talented, passionate and mission-oriented people! If you’re intrigued by our mission, check out our career page for open positions.