How did you come to be part of the imaging/machine vision industry?
It is amazing to think that I first started to get numbers out of images in 1964, a year before the word ‘pixel’ first appeared in print! I had specialised in metallurgy at Cambridge University, joined a local company called Metals Research in 1962, and the following year they developed the first ‘Quantitative Television Microscope’ for metallurgists in steelworks. This was an entirely analogue instrument – the results appeared on a moving coil meter and you used a slide rule to convert the output to real units. Eventually these came to be called image analysing computers and in the late 1970s industrial applications started to appear. I realised that there was a need to educate industry about how machine vision, as it became known, could be used to improve quality and enable automation, and in early 1983 I set up my independent consultancy to do just that, telling people from industry what was or wasn’t possible with current technology and if it was, who could do it.
In 1992 I helped to found the UK Industrial Vision Association as a not-for-profit trade association, and in 1995 I was asked to take over its administration, an activity that I am probably best known for today. In turn, the UKIVA helped to found the European Machine Vision Association in 2003, and I sit on its Executive Committee for UKIVA.
How do you convince customers that they need machine vision?
Historically, I convinced clients by explaining the technology and relevant methods, and introducing them to suppliers who had already done similar work. I always promised clients that I would tell them if vision was not appropriate, and I can recall one occasion when I said just that and suggested a much simpler solution. These days most of my work for end users of vision is free through the UKIVA, and we write dozens of articles for the industrial press giving examples of ‘success stories’ using vision; that is why we support the Machine Vision in European Manufacturing conference in April next year that is doing the same thing. In recent years my paid consultancy has been more for investors or about marketing innovations in vision.
What role does Europe have in the development of machine vision?
Compared with when I started my consultancy, Europe (and especially Germany) is hugely more important in the development of vision technology. A lot of the developments in CMOS cameras have come out of Europe, and there are probably many more makers of intelligent cameras in Europe than in North America.
What do you see as the major growth sectors?
'It is amazing to think that I first started to get numbers out of images in 1964, a year before the word "pixel" first appeared in print!'
Within the field of manufacturing, the reducing cost of systems, especially intelligent cameras, is going to make it worthwhile to inspect after each and every operation, avoiding adding value to an already defective product. In the wider context of machine vision there is a huge and almost untapped potential market for intelligent surveillance – store detectives can usually recognise actions seen on CCTV that precede any shoplifting act – but as far as I know you cannot yet buy a vision system with the same capability. Geographically, the so-called tiger economies of China and India must be major opportunities. At the European Machine Vision Association’s annual business conference in June, Professor Tieniu Tan, Director of the Institute of Automation of the Chinese Academy of Sciences in Beijing, told us that he had more than 100 PhD students studying vision at his institute! Personal experience with goods made in China suggests that better quality control is needed in some cases.
What do you see as the most important technological challenges facing the industry?
The number of highly trained people needed to install and maintain them currently limits the application of vision systems in industry. If vision systems were made as easy to use ‘out of the package’ as a mobile phone, a lot more would be used. Professor Jim Crowley gave an excellent and far-sighted exposition on this theme in his invited talk on ‘Autonomic computer vision systems’ in ICVS Bielefeld in March 2007, based on his own experience with a start-up company making vision systems to observe human activity – he found sales limited by the availability of engineers to install them.
What do you see as being the most significant commercial change in the industry during the years ahead?
The introduction of standards specific to the vision industry, such as Camera Link and GenICam, is going to make the component side of the industry much more of a commodity market, driving down prices even more. On the other hand, the company that can make their complete systems so robust that they can genuinely be applied in every case by the end user without any specialised training will win a huge market share. Although we are seeing consolidation in the industry, there seems no end to the appearance of new companies in vision.