Artificial Intelligence (AI) has found its usage in almost all various fields such as autonomous cars, predictive maintenance, facial recognition, text translation, satellite intelligence, and many more. GeoAI or Geospatial Artificial Intelligence is the combination of geographic information systems (GIS) and emerging technologies such as machine learning, deep learning, and AI. Geospatial experts deal with a lot of spatial data from Earth Observation satellites, SAR satellites, drones, aircraft, LiDAR data, location data, and so on. They are applying artificial intelligence algorithms to this spatial data to make data processing easier.
The Association of Geospatial Industries (AGI) co-hosted a webinar on the theme “Geospatial Technologies and Artificial Intelligence: Current Trends & Future” with the IIT Tirupati Navavishkar I-Hub Foundation.
IIT Tirupati Navavishkar I-Hub Foundation (IITTNiF) has hosted the Technology Innovation Hub (TIH) in Positioning and Precision Technology (PPT). The project is funded by the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS), Department of Science and Technology (DST), India. The project has funding of INR 100 crores over five years.
The Technology Innovation Hub (TIH) on PPT will provide a unique platform for researchers, industries, stakeholders, and end-users across multiple disciplines. The aim of TIH India is to align with the country’s broad objective of using GIS as an essential component for empowering Indian citizens and to become a top contributor to Government of India initiatives such as Make in India, Atmanirbhar Bharat Abhiyaan, and Start-up Ecosystem making India a self-reliant in PPT.
Existing Challenges
The webinar also addressed the problems in some areas where AI can be very useful to analyze data and help to make timely decisions. On positioning technologies, there are some areas where AI technology can be used for:
- Atomic clocks with improved time localization
- Navigation sensors for satellites
- Navigation sensors for Indoor mapping
- Quantum technologies for navigation
- Embedded hardware and IoT developments for various geospatial data applications
On precision technologies, geospatial data processing and applications are very complex, AI holds the power to simplify it.
Georg Hammerer, Chief Technology Officer, Hexagon Safety, Infrastructure & Geospatial covered the various aspects of using GeoAI to automate location intelligence. “From capturing the data by using GeoAI to produce results and put it into a solution, or dashboard including mobile or browser to make a better decision, GeoAI is very useful,” he adds.
Sanjiv Kumar Jha, Principal Smart Infra – SA, Amazon Web Services (AWS) talked about deep learning-based solution architectures for remote sensing image analysis.
He gave an example of urban analytics; “There are hundreds of use cases, 15+ domains, 50+ GIS layers, 50+ multiple integrations points to deal with such as mobility, transport, smart parking, citizen monitoring and many more. Location intelligence is only the keyword to generalize model,” he said.
Unique aspects of GeoAI
GeoAI has improved traditional geospatial analysis and mapping. It helps to understand and manage complex hyper-spectral data. Geospatial data is different and more complex from the data we used in AI:
- Data Type
- Data Volume
- Imagery Analytics
- Higher Dimensional Data
- Real-time applications
- Hyper-spectral
- Need for greater pre-processing
“Geospatial data cover hyperspectral data sets, to make decisions based on this data, we need high domain expertise,” said Saranya M, Assistant Manager – Presales, Esri India.
What can AI and ML do in geospatial data analysis?
- Extract images from imagery at scale, with high speed and accuracy
- Unstructured data can be converted into the structured format by using AI and ML
- Globally people are using AI and ML algorithms to update their base map and keep up to date
- In utilities; it helps in vegetation encroachment 2D, vegetation encroachment 3D, asset identification, and inspection.
- For security CCTV integration, traffic detection, crime protection, etc.
- In forestry tree count, assess tree health, detecting deforestation, catfish detection, etc.
Saranya M also discussed how AI/ML can be used to improve workflows. She talked about various applications of AI & ML models in geospatial, like pixel classification, object detection, instance segmentation, and image classification of floods or forest fires before and after.
