Machine vision in China

I’ve spent a lot of time looking at how best to go to market in China, including the competitive threats that will emerge there and how it is likely to happen.

I’ve spent a lot of time looking at how best to go to market in China, including the competitive threats that will emerge there and how it is likely to happen. I completely agree with your editorial (Vision Systems Design, January 2005, p. 88) stating that, to date, the relative niche nature of machine vision and the tremendous technical/application fragmentation have been a barrier to Sino-based competition. I would, however, like to add a few thoughts.

First, there are now machine-vision-system integrators in China. This is happening as a result of the first wave of precision electro-mechanical capital-equipment production in China and the establishment of sales and support arms for machine-vision component vendors from North America, Europe, and Japan. I would contend that much of the hardware in many machine-vision systems is relatively commodity: mass-market image sensors, standard camera ASICs, connectors, cables, DSP/ microprocessors, memory, etc. Most of these things are put into cameras and frame grabbers based largely on reference designs provided by IC manufacturers.

My feeling is that conditions exist for the Chinese to do the hardware side of machine vision whenever the confluence of interest, startup capital, and talent converge. These ingredients all exist there in isolation today, but it is just a matter of time.

The people, though, that have the expertise and application focus to pull it all together are the system integrators. These exist in China today, largely for the first time as it relates to machine vision. They are learning to work with the component technologies of major North American, European, and Japanese vendors. This group can handle the niche nature of machine vision and its fragmentation. With time, these integrators will want to expand their role, exploit local low-cost production capabilities, and, quite possibly, exploit lax IP protection in the PRC to assume a larger role in industrial machine vision.

Second, the driving applications for machine vision worldwide remain semiconductor, flat-panel display (FPD), and electronics capital equipment. There is a lot of talk about smart cameras breaking open heavy industry for machine vision. There is truth to this, but we’re still in the situation that as goes semi, FPD, and capital-equipment production, so goes the machine-vision industry. China is focusing tremendous national resources on the production of semiconductors and electronics. FPD is coming very soon. There are statistics about the large proportion of global production of electronics taking place already in China, and semiconductor production is becoming significant and very rapidly growing.

An indigenous capital-equipment capability for these sophisticated automation systems and instruments really does not exist today, with a few exceptions. Thus, the largest cohesive market that traditionally drives 50%-65% of machine-vision sales worldwide is still largely nonexistent in China. This will change as the Chinese seek to get more influence and control over critical supply chains for their most strategic industries, as well as expertise from the service and support arms of equipment producers who sell in China migrating to domestic pursuits.

This movement in the market chain should prompt considerable domestic activity in machine vision in China and create a foundation of competitiveness upon which to export abroad. I think that it is these emerging applications that will drive the biggest wave of Chinese-based competitors over the long term.

Dave Litwiller, Vice President
Dalsa Corp.

More in Image sensors