Podcast Audio Transcript
Alisha: Hello listeners, this is Alisha; thank you for tuning in to yet another exciting and informative podcast from us at Infosys BPM. Today, we are discussing about the transformative power of AI on spatial data. And to talk about this, we have here with us, Dr. Pradeep Kishore, Solution Design Head – Digital Interactive Services. Welcome, Pradeep. How are you?
Pradeep: I’m doing well, Alisha. Thank you for having me on this podcast.
Alisha: It’s a pleasure. Pradeep, spatial data is so crucial for several industries. The advent of AI and machine learning has transformed geographic information systems. Could you describe what challenges you see in figuring out GIS data?
Pradeep: Yes, Alisha. There are certain challenges in processing GIS data. GIS systems store geospatial data, relying on a large set of structured and unstructured map-related data. To gain precision, GIS systems need to rely on various sources, including satellite imagery; aerial photography; and LiDAR, which is light detection and ranging. They may also use more rudimentary sources such as paper maps, sketches, schematics, measurements, and survey data.
Since such varied sources are used, there is a lot of manual intervention and labour. This can in turn lead to data inaccuracies, due to the challenges in interpreting complex geometry diagrams involving engineering drawings, schematics, annotations, text, contours, and metadata.
Even though image recognition technologies such as OCR can help in the automatic reading of these images, they still require significant manual intervention to ensure quality in conversion, migration, correction, and cleansing.
Alisha: Those are some significant challenges you have to work with.
Pradeep: Indeed. With the advent of new technologies such as AI/ML and automation, you might need to become truly data-driven. Depending on the organization, this might mean eliminating legacy technologies, transitioning from outdated process models, changing outsourcing methodologies, and upskilling on digital technologies.
Alisha: So, with new technologies, there are also some housekeeping activities that need to be undertaken. Could you touch upon how you overcome these challenges?
Pradeep: Yes, to overcome these challenges, traditional GIS should be augmented with artificial intelligence and machine learning. The benefit of this is that it combines innovations in spatial science, AI/ML, data mining, and high-performance computing that are capable of extracting knowledge from spatial big data.
Alisha, it’s interesting to note that this augmentation of GIS with AI is one of the most interesting emerging innovations in the government and private sectors.
Alisha: That’s very futuristic indeed.
Could you touch upon some practical applications of using AI on spatial data?
Pradeep: Due to the rise of location-based systems and the cloud, spatial data has increased exponentially in recent times. AI/ML can be used effectively to tackle this spatial big data. AI/ML enables organizations to unlock the potential of data and deliver massive disruptions in geospatial data management. AI/ML can help organizations across various aspects, such as data workflow orchestration, information retrieval, business rules and automation, and operational & business analytics.
Another area in which AI can deliver great value is object detection, which is the process of locating and identifying assets. This also includes anomaly detection and feature extraction. This is very important in GIS as it involves finding objects on satellite imagery or aerial photographs and then plotting their location on a map. AI allows improvement in auto-extraction capabilities by almost 80%, while maintaining high quality and accuracy.
As an example, we enabled automated feature classification based on machine learning for a leading mining client. This made their entire mining site autonomous and helped extract features to build training data for their ML platform, in order to detect objects (such as equipment, field personnel, and structures) and to auto-pilot fleets of trucks.
Alisha: This quality of data is one of the things that come to mind when you mentioned AI/ML. Could you give some use cases where AI/ML can provide superior data quality?
Pradeep: Yes, as you mentioned data quality is extremely important. This is why considerable attention is paid to maintain data quality standards. Inaccuracies can be quite disruptive. Hence, traditionally we had labour-intensive and repetitive activities to maintain data quality. Also, with exponential data increase, this becomes even more arduous.
We can train a machine learning-based model to review source data or transaction data based on a predefined data model and produce desired results with reduced human intervention.
As an example, GIS specialists in utility industries can use ML to detect a missing valve type that needs to be added while assessing the diameter of connecting pipes and apply algorithms to highlight the same.
Let me give another example. Our client had transmission engineering data that had to be converted to the GIS format. It required a reconciliation of all entries through a manual process that led to errors. In this case, we built the annotation data and trained an ML model to scan information on the drawings within hours and produce a report showing the frequency of updates. An AI could then update the master database using this information, without losing quality.
Alisha: That’s very interesting. So, in your opinion, how will the future of GIS look like?
Pradeep: That’s a good question, Alisha.
A vast majority of data around us is geospatial in nature. In order to process this ever-increasing quantity of data, major AI advancements need to be made. AI/ML continues to create a technological revolution that is helping organizations offer innovative solutions to real-world problems. The accelerated progress and diffusion of new business innovations will also bring great transformation to our society.
Alisha: Sounds like a bright future to me. Pradeep, thank you for your valuable insights into this topic.
Pradeep: It was a pleasure, Alisha.
Alisha:Dear listeners, if you enjoyed our podcast today, please don’t forget to share and like it on social media. Our social handles are mentioned on the podcast page. The podcast will be available on various platforms like Google Podcasts and Spotify, in addition to our website.
Also, if you have any queries, do reach out to us through the email address on the podcast description. Watch this space for more exciting podcasts coming up. Once again, thank you for tuning in, stay safe, sharp, and healthy. Have a nice day!