/

December 22, 2021

Recapping 2021 and revealing what the future holds for Picterra

A recap by Frank de Morsier, Co-Founder & CTO

What a year it has been! As 2021 draws to a close, we thought it would be a good time to reflect on what the year has brought us, both from a product and company perspective. We recently asked our LinkedIn followers what they wanted our end-of-the-year blog post to focus on and their wish is our command!

We started the year getting ready to settle into our post-pandemic lives but quickly realized it probably wouldn’t be that easy. I am proud of how Picterrians powered through the uncertainty by adjusting to a hybrid work model, showing great empathy through team members’ health scares, and maintaining a fun, collaborative spirit throughout. We made it through a challenging year without losing sight of our goals!

A product strategy focused on delivering even more value to enterprise-scale teams

In 2021, we wanted to bring the extraction of Earth Observation information into the hands of large teams of GIS experts, data scientists, and business analysts. To accomplish this, we launched several new features targeted to enterprise-scale teams including:

  • Multi-class object detection: Instead of annotating a single type of object per detector, multi-class allows you to add different object/segmentation types to a single detector and annotate all of them in the same workflow. You can read about the benefits here.
  • Multispectral & non-visible imagery (e.g. SAR): Picterra supports and makes it easy to work with many different images types including multispectral and non-visible imagery. This capability is becoming increasingly important as satellite constellations continue to grow with new satellites, and their capabilities become more advanced. For example, Planet foresees the next generation of their Dove satellites to have 8-bands compared to the current 4-band classic setup (RGB + NIR). Synthetic Aperture Radar (SAR) imagery is becoming more accessible with constellations like Iceye or Capella space opening many possibilities to see through clouds.

     

    Multispectral

  • User management & permission control: We launched new organizational features to enable team collaboration and protect sensitive data. Admins have full control over user profiles and project permission sets across the account. Users can set view, edit, or manage permission levels for projects they own.
  • Interactive change reports: With a few clicks, Picterra users can present their geospatial analysis in a powerful change report and share it with stakeholders. This is particularly useful when comparing images from separate dates to understand what has changed and to identify patterns. You can test a demo report here.

  • Training reports: At the core of Picterra is the idea of iterating quickly on the creation of a model between the inputs given by the human expert and the model learning and providing feedback (visual in testing areas and with a score on accuracy areas) to help improve its performance.

    The platform now provides a

    “training report” with details on the model’s performance and what is happening under the hood. The user will also find recommendations on how to avoid common pitfalls when training the model. We aim to integrate more aspects of eXplainable AI (XAI) into the platform. The goal is to move away from a “black box” model and provide human experts with elements about how the model works and how it will perform on unseen data.

    Training report

  • Live collaboration: Working on a model to extract information is often a team effort requiring different sets of expertise. We are building the foundation for enabling live collaboration on a detector in the same way you would work seamlessly together on a Google doc! You are now able to keep working on your detector while it’s training. Soon colleagues within the same organization will be able to access it at the same time, too!

Building a dream team for even more expansive growth

But a company is nothing without its people. We are delighted to have added many new Picterrians to the tribe this year. Based in the Netherlands, we welcomed our new VP of Sales Pascal van Dalen. He has been instrumental in gearing up the commercial team and making Picterra a market hit. And hiring sales rockstars like Joy Bansingh, Account Manager, Antonius “Tim” de Weerd, SDR, and Samantha Bilski-Madiou, SDR!

In our Swiss HQ, we welcomed Viktoria Lindback, Product Marketing Manager, and Florencia Vassallo, Marketing Intern, who make sure Picterra gets noticed in the right places at the right time. Lara Engelien and Stefano Vojinovic, Customer Success Manager and Engineer, joined us to support clients through their onboarding and deliver the best possible support experience. Thank you all for joining us on this journey!

Despite periodic restrictions and lockdowns, the team found a few opportunities to get together outside the office to connect and laugh together. For example, we organized a hike in the beautiful Swiss Alp region of Valais and an excursion to the local escape game room. We hope 2022 will bring even more opportunities for team-building activities!

Team photo

Looking to the future: what does 2022 have in store for Picterra?

This has been an exciting year for Picterra. We are ending the year with a great mix of SME users as well as large-scale enterprise clients across sectors such as FMCG, ESG services, land mapping, civil engineering, and linear infrastructure. We have proven our thesis that there are many hidden application areas for geospatial intelligence that bring value to businesses. The sky is (literally) the limit for what our customers can achieve by having an easy way to extract insights from Earth Observation imagery.

On the product side, next year’s roadmap is centered around getting deep into collaboration. Think Google docs for geospatial imagery! Additionally, we will prioritize Active Learning – the principle of the model informing you with automatic recommendations for how to improve its learning. 

Moreover, our (multi-)human-in-the-loop approach is at the core of what we develop. It will further reduce the time required to develop robust machine learning models which is one of the most important metrics for ML teams. We will also invest in building out even more powerful quality assurance and control aspects to make large-scale analysis a no-brainer.

2022 promises to be a great year for Picterra. The company stands on a solid foundation, the platform is more powerful than ever, and we are expanding our dream team of growth-minded people. We are ready to hit the gas pedal! Soon we will announce our next big leap which will allow us to expand our customer base and go after even more use cases. We can’t wait to share more on that later.

Thanks to all our customers, employees, investors, and families for your support throughout this challenging, yet exciting year. We couldn’t have come this far without you!

We wish you all a peaceful holiday season. 

Want to learn more?

sign up to our newsletter