There is chatter aplenty on Linkedin and other media sources about Artificial Intelligence (AI) and its application in the drone-based inspection and data analysis market. The purpose of this blog is to dispel some of the myths and shine a bit of light on the realities of AI.
I am the co-founder of Sky-Futures but my background is not technology. My goal is to translate the technology we develop into a message that everyone can understand. Hopefully, this and future blogs can assist enterprises and others to sift through the marketing froth and get a true picture of what technology exists today and how it is useful to them. Chris Blackford

COO & Co-founder, Sky-Futures

What is Artificial Intelligence (AI)?

If you look up the meaning of AI, one way of describing it is, “…intelligence demonstrated by machines in contrast to the natural intelligence displayed by humans and other animals.” It applies where a machine mimics the cognitive functions that we associate with things such as human learning. However, people often misappropriate the term AI when describing automatic data analysis, data quantification or a computer vision process or technique. That isn’t to say that some of the techniques and processes used could not be described as AI, it’s just that the term is very overused when companies market their products.

What technology exists today to support automatic data/image analysis?

For me, technology is binary in terms of its application; it either adds value or it does not. There is a tendency to get about new technology because it can be as magical as Harry Potter. Once the excitement has subsided, there is one question that every enterprise must ask. Is the technology relevant and does it add value?

At Sky-Futures, we are focused on the industrial inspection market. Think large, vertical, metal structures and the like. In these markets, drones are used to inspect cell towers, transmission towers, oil rigs, ships, bridges etc. The inspection process is looking for anomalies such as corrosion, cracks, bends, breaks and missing parts.

Our goal at Sky-Futures is to bring a fully automated, end-to-end, drone-based inspection solution to clients through our Expanse software platform. Whilst we cannot directly influence how drone hardware develops, we have total control over how and where we automate the data analysis and data reporting process. The technology we have developed (and continue to develop), solves real problems faced by real customers.

I’m now going to bring this blog back to AI, or rather the early stages of what is often referred to as AI. Automating the entire data analysis and reporting process will not happen overnight. It requires large volumes of data and the right tools and techniques being applied to that data. At Sky-Futures, we are constantly bringing new technology into our services that deliver value today rather than something sentient that is fantastical and still years away. You only have to search for AI online to see the plethora of videos promoting something that looks incredible but won’t exist for years. The technology we sell already exists. It works and is in use by customers all over the world – in the areas of Renewables, Oil and Gas, Utilities, Telecoms and Bridges, and more.

A recent example of the application of our technology comes from the electricity transmission and distribution (T&D) market. Customers can have thousands or tens of thousands of towers. Traditional inspection techniques rely on climbers or helicopter inspections collecting large volumes of image data packaged up with…no software and just the human eye. Imagine wading through thousands of high MP imagery to find anomalies and report on the condition of the towers. A slow and extremely tedious process that likely leads to a lot of information being missed. It is also an incredibly inefficient use of an engineer’s time.

Sky-Futures partners with Inspection² to remove this inefficiency and turn engineers from being report writers to decision makers. For the T&D market, one of our machine learning tools will automatically locate corrosion on transmission towers. If you are a company with 1,000 towers, your engineers are faced with inspecting thousands of images to find, locate and grade corrosion. Why leave an engineer to spend days poring over this amount of data when a machine learning tool could do it in less than an hour? This makes the process many times more efficient and helps focus the engineer on the data that allows them to make asset integrity and operational decisions.

This is one of many examples of Sky-Futures using machine learning and computer vision tools to significantly enhance the drone-based inspection and data analysis process.

Share This