What does a machine vision inspection system consist of?

Machine vision is the simplest introduction: using machines to replace human eyes in work. Machine vision is not an independent entity; it is a complex and precise system. The machine vision system captures image information through visual imaging devices and then transmits it to a dedicated image processing system. According to the information set in advance in the system, such as pixel distribution, brightness, and color, it is transformed into digital signals. The image system then performs various operations on these signals to extract target features for recognition, positioning, and judgment. Based on the operation results, the equipment on-site is controlled to perform a series of operations.

So, what does a machine vision system consist of?

1.Illumination light source

The light source in machine vision can be classified according to shape, light emission principle, and spectral characteristics, but the most core is still the light emission principle. According to the light emission principle, there are: LED light source, high-frequency fluorescent lamp, halogen lamp, and laser light source. These light sources have their own advantages and disadvantages:

• LED light source advantages: long service life, high brightness, fast response speed, can be customized according to requirements for different wavelengths, color temperatures, and color rendering indices.
• Fiber halogen lamp source advantages: high brightness, almost no change in brightness and color temperature. Disadvantages: slow response speed.
• Xenon lamp source advantages: high brightness, color temperature close to sunlight. Disadvantages: large heat generation, short life.
• Fluorescent lamp advantages: good diffusion, suitable for large-area uniform illumination. Disadvantages: slow response speed.
• Laser light source advantages: provides extremely pure and uniform light, generally used for high-precision measurement and detection. Disadvantages: high cost.

2.Lens

The role of the lens is similar to the human eye lens, adjusting the size of the light beam. In the machine vision system, the main function of the lens is to collect and transmit the light from the imaging target to the CMOS chip. The better the light transmission of the lens, the more light it provides to the CMOS chip, and the clearer the photo will be. Therefore, the quality of the lens directly affects the overall performance of the machine vision system, so choosing a good lens is an important part of the machine vision system.

3.Industrial camera

The industrial camera is a subcategory of cameras and is also one of the most important components of the machine vision system. The function of the industrial camera is to convert the light signal into an orderly electrical signal and transmit the signal to the computer for image capture and storage. Compared with consumer cameras, industrial cameras require higher image stability, faster transmission capabilities, and anti-interference capabilities that consumer cameras do not have. The most important part of the industrial camera is the image sensor, which is divided into CCD and CMOS, so industrial cameras are also divided into CCD industrial cameras and CMOS industrial cameras.

CCD was the most commonly used image sensor in early industrial cameras. It has the advantages of high-quality images, low noise, high sensitivity, and wide dynamic range, but has the disadvantages of high power consumption, large volume, large heat generation, and most importantly, high cost. This was also the biggest reason why the machine vision system could not be popularized at the time.

CMOS image sensors first appeared in the 1970s. In the 1990s, due to the maturity of ultra-large-scale integrated circuit manufacturing technology, people integrated the photo-sensitive element array, image signal amplifier, signal reading circuit, analog-to-digital conversion circuit, image signal processor, and controller on a CMOS chip, making CMOS’s image quality no worse than CCD. In addition, CMOS industrial cameras have lower power consumption, higher frame rates, better integration, and are more compatible with standard silicon-based microelectronics manufacturing processes. The cost of CMOS production is far lower than that of CCD. Better image quality and lower cost have made CMOS industrial cameras successfully replace CCD industrial cameras and become the mainstream in the industrial camera market. At the same time, it has also made machine vision a product that everyone can afford.

4.Image transmission interface

Image transmission is an aspect of machine vision that is easily overlooked, but its importance is no less than that of the lens. If the image transmission is not efficient during the transmission process or is interfered with by other electromagnetic signals, the backend computer will not be able to receive the picture, or the received picture will be incomplete, causing the entire machine vision system to stop working or make serious operational mistakes. The factors leading to this situation are very complex in the industrial application environment. Machine vision manufacturers have developed various transmission protocol interfaces, such as Gige interface, USB2.0/3.0 interface, CXP interface, optical fiber interface, cameralink interface, etc., to meet the requirements of image transmission efficiency, transmission distance, anti-interference performance, and cost. While ensuring the stability and reliability of image transmission, it also provides a more cost-effective choice.

5.Image processing system

The processing of image information in the machine vision system includes both hardware and software. They are the key to machine vision replacing manual inspection and are the brain of machine vision.

• Image processing hardware:Image processing hardware refers to computers or industrial control machines that can receive image information from industrial cameras and can run image processing analysis software. It is the information carrier and processing center of the entire machine vision.
• Image processing software:After the image information captured by the industrial camera is transmitted to the computer, it needs to be processed by software including image processing algorithms, pattern recognition algorithms, machine learning algorithms, etc., for image enhancement, data encoding, edge sharpening, segmentation, feature extraction, and other image processing. Then, the correct operation instructions are output to let mechanical parts such as robot arms, conveyor belts, etc., perform specified work. If the industrial camera, light source, lens, computer, and image transmission interface are equivalent to the entire human system, then the image processing software is the soul of machine vision.

In the choice between machine vision and manual inspection, for the quality control department of factories, they would prefer machine vision with better efficiency, precision, and execution, and without any complaints.

From a cost perspective, over a long period, machine vision is definitely cheaper than manual labor. Quality inspection is a necessary position in every factory, and the annual cost of four quality inspectors is generally more than 200,000. Although the price of machine vision systems varies across industries, a budget of 200,000 is also very sufficient to buy a set of machine vision systems. Considering cost, work efficiency, and quality, no factory would refuse a machine vision system.

In the short term, machine vision cannot completely replace manual visual inspection, and some more complex scenarios cannot be met due to technical reasons. However, with the development of AI technology and the continuous enhancement of AI learning capabilities, it is only a matter of time before machine vision systems replace manual visual inspection. This is the trend of the future and the wave of the times.