Computer Vision can uncover several nuances of the visual environment that humans cannot, and it is this capability that is a real game changer !

Shreedhar Rangappa (PhD), our next pathbreaker, Senior Research And Development Engineer at Honeywell, works on solutions to challenges in the aerospace and avionics industry using computer vision approaches.

Shreedhar talks to Shyam Krishnamurthy from The Interview Portal about his internship project on optical 3D metrology that was a pivotal point that led him to a career in computational imaging.

For students, focus on the difficult side of the story rather than the easier part, because the tougher challenges bring out the best in you !

Shreedhar, Your background?

I grew up in a town named Tumkur in Karnataka. My family consists of my dad, mom, and elder sister. Dad worked in the army and later in the Banking sector. Hence, discipline and punctuality were the two thumb rules. 

From a young age, I was interested in building things from scratch, rather than playing with readymade toys. Hence my parents believed that I had an interest in Engineering compared to other career opportunities out there. I was an active member of the science society at school along with other activities in Scouts-guides, NCC and sports.

What did you do for graduation/post graduation?

I studied Electrical and Electronics during graduation and Mechatronics systems during my Masters. I went for Computational imaging as my area of focus during my PhD.  

What were some of the influences that led you to such an offbeat, unconventional and uncommon career?

My interest in the field was the key influencer. My parents, especially my dad and I, did all kinds of household jobs like carpentering, plumbing and mechanical work, which ignited the spark in me to pursue engineering as my lifeline.

My 8th-grade science teacher was an awesome person who did not teach the course from the book but rather from the questions and answers that he discussed in the class. Hence, students tend to forget the subject, not the discussion.

Before I got into engineering, I was particular about my study patterns. I prepared for every class and not just for exams or internals. I have always been a believer, and tried to prepare for the worst. Preparing mentally for the worst gives ample time and strength to deal with anything lesser than the worst. 

I strived hard to get a job before I could finish my degree, and I got into TCS as an Associate Software Engineer. Within a few years, I realized that there was not much of a challenge in my job. To keep my hunger for learning alive, I started preparing for higher studies and got into Kingston University in the UK for my masters in Mechatronics/Robotics. 

I graduated from my masters with an internship at National Physical Laboratories (NPL) in the UK.

The internship was on testing and validation of an optical metrology instrument called Alicona. This instrument works on focus variation technology, which is a precise handshake between optics and electronics. And this is where the field of study, Mechatronics, made sense. I was able to apply the theoretical knowledge I gained at the university during my internship. I was validating how well the instrument was able to measure the surfaces (measuring how smooth a surface is) and at what critical angle the instrument can fail to measure. I was introduced to the field of optics, electronics, mechanics, and metrology during my internship. The main idea behind this work was to get more information on generating 3D information on a metal surface with a single camera. 

This was the pivotal point that helped me grab a fully funded PhD scholarship at Loughborough University, UK to pursue a career in the field of Optical Metrology and Computational Imaging. 

Today, 3D depth reconstruction needs either 2 cameras or more single cameras in multiple positions to get depth information to an extent that is usable.  During the internship, the 3d information was generated by changing the camera focus and measuring the focus change in samples.

How did you plan the steps to get into the career you wanted? Or how did you make a transition to a new career? Tell us about your career path

During my internship, I was exposed to the method of generating 3d depth information, by moving the camera focus rings. Yet, there were some physical movements involved.

My PhD topic was to investigate the possibility of a movement-less 3D depth-generating camera and its uses in 3D metrology. I used a new type of lens called Micro-lens arrays that resembles a “Bee eye”. With this new lens combination, the depth could be detected and the focus zone could be moved using the software. The micro-lens combinations were computationally arranged to get a picture. Some of the cons of this technology were the lower resolution. Also, the depth range was not high. Yet the technology could be used for many applications.

With the knowledge gained, I started working as Research Associate with the sustainable manufacturing team. Here I worked on automating human and water-intensive tasks in the food manufacturing industry. Water intensive tasks involved huge containers that needed to be cleaned in a timely manner in the food industry. These were water and time guzzling tasks. With a camera illuminated by UV light, the cleaning score was generated using classical and machine learning methods.

Next, I moved into the analytics field in the sports industry. In European countries, football is an important game, and there were many startups working on player analytics to improve their performance and making changes to win. We developed new player-tracking algorithms, player-detection methods and later on, analytics on each player. We collaborated with county clubs to get rich datasets and an automation tool for major annotation tasks. Prior to this, there was a large team of humans who would go through the football video and get the metrics of each player. This was taking around a week before the coach could help individuals with their performance and key areas of improvement. This manual process was not providing players an understanding of their weak areas and getting enough training before the next match. 

With the new technology and cloud computing, coaches could help individuals within a few hours of the actual match, with all key analytics generated automatically.

I returned back to India, venturing into a new field to apply my skills and that’s when I met the Pace Robotics team. I was the initial co-founder and took care of the image processing division. 

Here, the idea was to automate the labour intensive task of house construction. Initially, we started with automation of plastering walls and painting. 

During my PhD, I realized the difference between a PhD holder and a normal student. A PhD journey teaches lessons the hard way with ample time to learn from them. A PhD aspirant would not accept something that won’t work in a couple of attempts, rather he/she tries and makes all possible efforts to make it happen. This lesson has proved itself right many times in my life.

How did you get your first break?

I was finishing an internship at NPL, and some of the rules changed in the UK that allow foreign students to get jobs. That is when I realized that I need to get more than a job in the UK and started looking for scholarships for a PhD. After applying to numerous institutions, I got an opportunity at Loughborough University, where I was about to get a chance to learn more than I could consume in all fields, especially computer vision, machine learning and 3D-Metrology.

During this period, I made good contacts back in Bangalore to work on side projects in the field of Sericulture, which would pave the way for more research-oriented work in upcoming years.

What were some of the challenges you worked on? How did you address them?

I will talk about some of the challenges I faced during my work.

Challenge 1: 

There has been a more economical loss to sericulture farmers due to a lack of technology penetration. The cocoon yield depends on various factors and the harvest they want to sell at the market does not provide the price they are looking for. The middleman would give a price quote for the cocoon which is a lot lesser than what the seller would buy. If the quote is at a lesser rate, the farmer would suffer loss, just because of a few damaged cocoons in his lot.

In order to digitize the sericulture industry, we introduced many products, at different levels of the silkworm life cycle. The eggs laid by silkworms are very tiny and take a lot of effort for a human to count. The egg count directly impacts the final price/rate of the silk at the market and hence accurate egg count becomes essential. We build a computer vision model aided by AI to count and classify the eggs in a few minutes, which would take more than 25 mins for human workers.

Where do you work now? What problems do you solve?

I work as Research and development Engineer at Honeywell and I try to provide solutions to the aerospace and avionics industry using computer vision.

At Honeywell, we have diversified support for passenger aircraft throughout their journey from land to air and back to land. We have a lot of computer vision applications within the cockpit helping pilots for smooth journeys in the air. I currently take care of landing the aircraft safely at the correct airport gate (I cannot go into details as they are customer-specific details)

What skills are needed for your role? How did you acquire the skills?

A degree in the field of CS/EEE/EC would be ideal for the role. Since I have EEE and PhD in computer vision, that helps me to do well in my current role

What’s a typical day like?

A typical day would be spent in a lot of discussions related to a specific problem in avionics. It would involve brainstorming all possible ways to provide the solution to a particular problem using computer vision. Once we have settled with an idea, it is programmed as a solution. The iterative process of development, testing and tuning would be the task on a given day  

What is it you love about this job? 

I love the nature of research backed solutions that we provide to the customer.

How does your work benefit society? 

I am contributing to society so they can have a safe and enjoyable journey through air. The other research is helping sericulture farmers utilize technology in farming to predict their harvest and yield before they could see it.

Tell us an example of a specific memorable work you did that is very close to you!

The sericulture project that I worked on, gave me an opportunity to meet new people from the rural side of the country and give the best to them in terms of technology products. 

When sericulture farmers get the right yield from their cocoons, and they are happy, that gives a true meaning to my work.

Your advice to students based on your experience?

I would urge the younger generation to focus on the difficult side of the story rather than the easy part. Try to get to the roots of the problem. Fix the problem, rather than just trying to find a workaround.

Future Plans?

I would like to continue working in the avionics field, where there are a lot of opportunities for computer vision and AI. I would also spend some time coming up with innovative solutions that make Indian farmers ( sericulture) enjoy their work in the field rather than struggling to make ends meet.