Machine vision is one of the technologies that realizes factory automation. The machine vision market has changed significantly in the last 10 years. In this series, we will explain the changes that have occurred in the machine vision market over the last 10 years, current trends, and future prospects.
In recent years, as image processing technology has been applied to a wide range of fields, an overwhelming number of cameras, which are input devices for image processing, have become widespread in the market compared to a few years ago.
It is well known that the cameras installed in automobiles and smartphones (smartphones) have been a major driver of this trend.
From a technical point of view, the change from CCD to CMOS in image sensors has made this trend possible by increasing sensor productivity, achieving low cost, low power consumption, and low size. Of course, because it was used in consumer products such as automobiles and smartphones, the price dropped due to mass production, which also contributed to its widespread use in the market. It can be said that the improvement in technical productivity and adaptation in the consumer sector have created a great swell.
The spread of cameras will continue to accelerate. Looking only at automobiles, in addition to the back monitor and drive recorder that are becoming standard, collision avoidance and automatic parking will become standard in the future, and further, automatic driving, driver’s doze detection, and image sensing technology will be required. The field is expanding. Similarly, automatic driving is not limited to automobiles, but cameras are beginning to be installed in automatic traveling forklifts, cleaning robots, drones, service robots, etc. at distribution centers.
Unlike the spread of cameras in our familiar world, there are areas where cameras have been used frequently even before that. That is image processing for FA applications that helps factory automation, and is commonly referred to as machine vision. It is a technology that automates manufacturing in place of the human eye and brain, and its applications include inspection (defects, foreign substances, cracks, cracks, stains, etc.), position detection, measurement, number counting, barcode reading, character recognition, etc. Bulk picking and the like can be mentioned. The fields of application are wide-ranging, such as semiconductor manufacturing equipment, electrical and electronic equipment, robots, automobiles, medical care (pharmaceuticals), printing, logistics, and agriculture.
Cameras required in the field of machine vision are called industrial cameras, and are different from ordinary consumer single-lens reflex cameras, digital cameras, smartphones, and cameras equipped with cleaning robots. The major differences are:
- Long-term supply
- high resolution
- Fast capture
- Rugged housing: Imaging with a trigger signal from an external device
Of these, the most important is long-term supply. For example, a single-lens reflex camera can obtain very high-quality images, but due to frequent model changes, it is not possible to purchase the same model continuously for three years or more. When the model is changed, it requires redesign, development, and verification of the entire device including lenses, lighting, software processing, etc., which requires enormous man-hours. In addition, cameras such as smartphones cannot obtain image quality that can withstand industrial use, and cannot be used for inspections. Therefore, consumer cameras and industrial cameras are completely different fields, and the active players (manufacturers) are also different.
Until now, industrial cameras have been used in factories around the world. However, from around 2005, overseas manufacturers began to take the share of the world at once, and the market share changed completely in the last 10 years or so.
