Home Articles Petabytes of geospatial data to bite-sized actionable insights

Petabytes of geospatial data to bite-sized actionable insights

Technological advances and innovations in the recent past have acted as an enabler for our business, starting with the advancement in Deep Learning. The novel architecture supported by CNN, RNN algorithms have aided in solving complex computer vision problems.

With the rise of cloud services such as Azure, AWS and, GCP, we have come a long way from the time when satellite imagery was shared in a physical hard disk. Today, we can stream it directly from the cloud to our desktops. Leveraging the auto-scaling infrastructure on the cloud helps us process large volumes of satellite imagery in a very short time.

Then there is the smallsat revolution — we now have satellites that are much smaller and weigh far less than the conventional school bus-sized satellites. The influx of satellite imagery on a near-daily basis combined with our computer vision technology has helped us unlock previously unimaginable applications. With Airbus’ Neo and Maxar’s Legion being launched in the coming years, high-cadence, high-resolution imagery is going to unlock endless possibilities from monitoring sites for security to tracking port movements for commodities to numerous other applications.

Also Read: Following geo-solutions business model

Digital journey

The world embarked on a digital journey a couple of decades ago, and we saw the PC, Internet, and Cloud completely changing the way businesses work. Today, Big Data & AI are driving the digitization process and HyperVerge is enabling various organizations to leverage the same. We build deep learning-based AI algorithms that transform petabytes of data (images, videos & documents) to bite-sized actionable and accurate insights in tactically relevant timelines for defense and commercial sectors.

Last year, we worked with DNVGL, an energy consulting company based out of Dallas and built a Digital Twin of the entire transmission energy infrastructure of Texas (695,662 km2) using high-resolution 30cm DigitalGlobe satellite imagery. We deployed our asset detection algorithms to identify all the substations, transmission towers, wind turbines, and solar farms with extremely high accuracy (>99%). We transformed and automated a manual task which took a few weeks to complete earlier, to conclude in a matter of minutes. The same technology has also been replicated to identify other assets of interest like roads, buildings, nuclear power plants, fighter jets, ships, etc.

We have worked across sectors such as O&G, defense, self-driving cars amongst many others to automate mission-critical tasks and offer best in class accuracy. Considering that 2 million commercial satellite images are being downlinked every day, processing this data in a short period of time is beyond human capacity. With our change detection algorithms, one can identify relevant changes across varied geographies and seasons e.g. changes along a border, illegal ship movement across water bodies, etc. Earlier if an analyst had to look at 100,000+ images to identify relevant changes, with our solution he has to look at only those 100 images where actual changes have happened, making the process of ingesting petabytes of data a matter of few seconds thereby surpassing human-level performance.

Also Read: 5G, data revolutionizing drone industry

Collaborative Innovation

Our founding team has been working on computer vision since 2009, when we were part of a student research group at IIT Madras. The amount of experience that the team brings to the table has helped us solve complex fundamental problems in an agile manner. This year, our focus will be on the defense, telecom, insurance and energy sectors. We believe that one of the most effective ways to accelerate the creation of value across sectors is to collaborate and leverage complementary capabilities. Hence, we are investing in our R&D by collaborating with various domain partners, satellite and aerial imagery companies. To fast forward our product development cycles, we have also been participating in a few industry-focused accelerators. In September 2019, we were one of the eight startups that were part of the Catalyst Space Accelerator, which is a defense and national security industry accelerator sponsored by US Air Force Research Laboratory in Colorado Springs.

HyperVerge was part of a collaborative ecosystem where the defense industry and Colorado’s aerospace community came together to help warfighters get tactical information in near real-time. We further plan to venture deep in this domain by enhancing our space-based intelligence, surveillance and reconnaissance capabilities, and help defense contractors, intelligence agencies and governments take appropriate measures for national security and public safety.

On the commercial front, we are one of the three companies which is currently a part of the Airbus BizLab accelerator in India, where HyperVerge and Airbus intrapreneurs are working together to speed-up the transformation of innovative ideas into valuable businesses, while exploring possible synergies. We plan to use this as a springboard for our venture into new geographies and sectors which show promise.

Growth opportunity in recession

The economic slowdown is unfortunate. However, it is also an opportunity for companies like us that offer new technologies. Since margins are getting smaller, there is a dire need for companies and governments to find new sources of revenues and cut costs at the same time. Non-commercial sectors like defense haven’t been affected much by the downturn. However, the impact is visible on commercial sectors like telecom and energy, which are dependent on the economic condition of the state. With the cost of adopting new technology soaring, it becomes a vicious cycle in which even Fortune 500 companies are apprehensive of investing time, effort and money.

To circumvent this, at HyperVerge, we have set up a low-cost structure by undertaking most of the technology development in emerging markets like India. This helps our clients to experiment and innovate, even during an economic downturn when they have a limited budget at hand. This has enabled us to grow aggressively despite unfavourable market conditions.

To share an example, most of the large telecom companies in India are facing tremendous pressure to adopt digital workflows & cut cost. Telecom operators like Reliance Jio, Vodafone and Idea have partnered with us to use AI to do a digital KYC instead of a manual one, saving over $5 per new customer. The savings allowed our clients to reallocate this money to other critical areas. Together they have added over 100 million customers in 2019 and HyperVerge has been a crucial part of this journey.

During the last recession, traditional software businesses with legacy software were disrupted by SaaS and Cloud infrastructure. Today, it is the age of agile AI, which will win over expensive setups.

Tips for survival

It is always prudent for companies to reduce wastage and increase productivity. In recessionary times, it’s mandatory — that’s the only way to survive the downturn. While the governments of the world have to provide the economic stimulus to revive growth, companies like ours can help by leveraging new technologies for collecting reliable data and monitoring recovery, while government measures are implemented.

Let’s say if the infrastructure industry is going through a turmoil and is known to be the most affected, the relief packages sent by the government would be crucial to start the engine again. Geospatial technologies can become a medium to help policymakers monitor infrastructure developments happening across the country, thereby ascertaining the outcome of the relief packages on a weekly to monthly basis. Similarly, in the case of retail, where consumer spending drops, policy changes in lending rates and providing various subsidies can enable a rise in demand. Government bodies can use anonymized GPS data from cell phones, monitor vehicle movements from imagery and overlay other open-source (geospatial) information like demographics on a map to better understand changes in footfall around retail stores and confirm if the changes in policy have been well received by the end consumers. In both cases, the bottom line is that the feedback loop of action-reaction is complete, aiding in quicker and agile decision-making.

Also Read: Turning Drones into Intelligent System