EOfactory.ai: Now everyone can apply AI in Earth Observation
From data acquisition and processing, all the way to building BI dashboards, we have come a long way from a small-scale project solution provider to a global platform. We support both small/medium commercial clients and large enterprises, including government institutions, in translating big data pixels into actionable intelligence, says EOfactory.ai CEO Abhay Mittal, in an interview with Geospatial World.
Can you tell us about EOfactory.ai and how it is different from other similar platforms in the market?
Data is everywhere and it is increasing every day at an exponential rate, especially with the introduction of new optical, hyperspectral, m-wave, GHG, aerial platforms, and drone sensors. With satellites, drones, and aerial platforms being launched in global markets almost every month, how could we contribute to this unprecedented amount of data going to the users? One known fact is that many geospatial experts are domain experts, not programing experts; hence, we envisioned a platform that could help these professionals maximize the benefits of Earth Observation with Remote Sensing + GIS + Artificial Intelligence (AI) and Business Intelligence (BI) dashboards.
The origin of EOfactory. ai’s concept goes back to the days when I tried using desktop-based GIS and image processing software to generate relevant insights for an intuitive and visually appealing dashboard. The challenge was to find the right icons to click in an ocean of menus and options to select, without having the option to display such results on an optimized dashboard.
Another challenge I faced was the expensive cost of acquiring such software. I saw this opportunity and gap in this market segment and launched EOfac- tory.ai as a unique virtual Earth Observation ‘factory’ on the Cloud connecting different stakeholders. With SAAS (software as a service) as the driver from the start, we built a powerful and scalable platform that allows for scaling up of processing. The goal was to bring in the state-of-the-art technology framework that everyone could access and integrate into their workflows.
We connect to multiple data sources and download them to shared workspaces which everyone can collaborate on to derive insights. We have built the architecture to support and run our AI/ML algorithms and processing at the data source to avoid unnecessary transfer of huge data volumes. It also protects the movement of data within the network.
We also allow users to build their own AI models and run them on our EOfactory.ai platform, which gives researchers the confidence of protecting their algorithms and models of their own research work. With EOfactory.ai, domain experts can focus on their domain-related workflows and use automation for some of the other repeatable tasks such as downloading imagery and preprocessing.
We built a time-series management framework with an integrated open data cube storage. Time series analytics together with Machine Learning capabilities allows us to support multiple workflows in sectors like agriculture, forestry, telecommunications, defense, etc. From data acquisition and processing all the way to building BI dashboards, we have come a long way from a small-scale project solution provider to a global platform supporting both small/medium commercial clients and large enterprises, including government institutions, in translating big data pixels into actionable intelligence. By operating on the Cloud, we are able to scale up and provide processing power and efficiency much faster than traditional software. Using Kubernetes, we are able to cluster and scale up micro-services for specific workflows.
Can you tell us about some of the sectors in which you are currently operating, and the areas you will be targeting in the near future?
We are currently focusing on agriculture and forestry sectors using both satellite and drone data. We have also been doing some research and development work for the construction and defense sectors. Although our platform caters to a wide range of industries, it is important that we stay focused on a couple of areas and deep dive into those sectors. Building workflows specifically to solve problems in the agriculture and forestry sectors allows us to deliver solutions at the national level. We have built a community framework under which we will be launching programs for the communities to contribute to the development of other custom workflows. Our technology team is composed of data scientists in AI, Remote Sensing, GIS, and BI to put it all together for our users. Any functionality that is required for a specific workflow can be built and delivered in weekly sprints in an agile framework methodology that is applied in the platform.
We derive insights into object detection and feature extraction as the two main drivers of the platform. Identifying multi-ob- jects such as solar panels, planes, ships, cars, swimming pools, trees and extracting features such as building footprints, land use, landcover, crops, etc. are the building blocks of the workflows and models we establish for clients to extract value in an automated way from the vast source of relevant data. Prediction and risk analysis are two key areas we are currently working on.
How does EOfactory provide customized solutions to different stakeholders in the agriculture sector?
EOfactory has developed its own workflows for the agri- culture sector by generating automated farm-level data, which is built using proprietary algorithms, and then generating monthly, fortnightly or weekly analytics on these farm-level data. We have also integrated data capture through field surveys to the display of analytics on the dashboard. Through this application and the use of geo-tagging, we are able to collect details of farmers, plots, crops, and their conditions. This information is connected to a web dashboard that provides a synoptic view of all field data that helps in staff management and monitoring. These solutions can be used by fertilizer, insurance, seed, or other agencies which require farm monitoring. We are also exploring workflows to increase security and transparency at the farm level by developing a framework for implementing largescale blockchain technology.