As various industries in my country begin to widely demand industrial automation and intelligence using image and machine vision technologies, machine vision is gradually being applied in industrial settings.
Driven by favorable government policies, China's machine vision industry is developing rapidly, and China is becoming one of the most dynamic regions in the world for machine vision development. It is estimated that by 2025, the market size of machine vision in my country will reach 24.6 billion yuan. Let's learn more about machine vision—the "eyes" of 5G industry.
Machine vision is a comprehensive technology that includes image processing, mechanical engineering, control, electric lighting, optical imaging, sensors, analog and digital video technology, and computer hardware and software technology (image enhancement and analysis algorithms, image cards, I/O cards, etc.).
A typical machine vision application system includes image capture, a light source system, an image digitization module, a digital image processing module, an intelligent judgment and decision-making module, and a mechanical control and execution module.
In summary, machine vision is a comprehensive technology combining software and hardware. It requires software to process images, and hardware to provide stable and high-quality images; both are equally important. As a developer, I am more familiar with the software modules, but my knowledge of hardware systems is weaker. This article provides a simple summary of hardware-related resources in machine vision as a reference for beginners.
Content Summary
1. Industrial PC
An industrial PC can be understood as a commonly used PC host, but it is more powerful in image acquisition and processing, as well as related control and interface capabilities. In a machine vision system, the performance of the industrial PC directly affects the processing speed and runtime of the entire vision system, making it a critical component. The selection of an industrial PC needs to be considered from the following four aspects:
1) Dimensions
2) Installation method
3) Configuration
Choose the appropriate configuration based on your application scenario.
4) Interface
This is very important; it determines the number of cameras, light source controllers, and connection methods in the vision system. For example, if the system has 8 cameras, 4 light sources to be controlled, the cameras are connected via network cables, and the light source controllers use a RS-232 interface, then the number of network interfaces and serial ports on the industrial control computer can be determined.
2. Camera
2.1 Camera Type
Dot scan camera / area scan camera
Area scan cameras: These achieve pixel matrix shooting. In a camera's image capture, the detail is determined not by the number of pixels, but by resolution. Resolution is determined by the focal length of the lens used; the same camera will have different resolutions depending on the lens's focal length. Since the number of pixels does not determine image resolution (sharpness), what are the advantages of a high-pixel camera? The answer is simple: fewer shots and faster testing.
Line scan camera: As the name suggests, it is in the form of a "line". Although it is also a two-dimensional image, it is extremely long, several kilobytes in length, while its width is only a few pixels. Generally, this type of camera is only used in two situations: 1. When the field of view to be measured is a long and narrow strip, it is often used for problems such as inspection on rollers.
Second, a very large field of view or extremely high precision is required. In the second case (requiring a very large field of view or extremely high precision), it is necessary to use an excitation device to repeatedly excite the camera, take multiple pictures, and then merge the multiple "strip" images into a single large image.
Therefore, using a line scan camera requires a capture card that supports it. Line scan cameras are expensive, and their detection speed is slow in situations requiring a large field of view or high precision—typically, a camera's image is 400KB to 1MB, while the merged image is several megabytes in size, naturally slowing down the process. Slow and steady wins the race. For these two reasons, line scan cameras are only used in very special circumstances.
CCD camera / CMOS camera
CCD cameras offer excellent image quality and noise reduction. Although the addition of external circuitry increases system size and reduces reproduction costs, it also provides circuit designers with greater flexibility to enhance specific performance aspects of CCD cameras. CCDs are better suited for applications with very high camera performance requirements but less stringent cost control, such as astronomy, high-resolution medical X-ray imaging, and other applications requiring long exposures and stringent image noise control.
CMOS cameras offer advantages such as high yield, high integration, low power consumption, and low price. However, they inherently produce more image noise. Current CMOS technology has continuously evolved, overcoming many of its early shortcomings and reaching a level of image quality comparable to CCD technology.
CMOS is suitable for applications requiring small space, small size, and low power consumption, where image noise and quality requirements are not particularly high. Examples include most industrial inspection applications with auxiliary lighting, security applications, and most consumer digital cameras. Currently, CCD industrial cameras still dominate vision inspection solutions.
2.2 Resolution
The number of pixels a camera captures in each image generally corresponds to the number of pixels arranged on the target surface of the photoelectric sensor. The choice of resolution also depends on the application scenario and accuracy requirements; higher resolution is not always better.
2.3 Pixel Depth
The number of bits per pixel is commonly 8-bit, 10-bit, or 12-bit. Resolution and pixel depth together determine the image size. For example, a 5-megapixel image with an 8-bit pixel depth would be 5 million * 8 / 1024 / 1024 = 37MB (1024 bytes = 1KB, 1024KB = 1MB). Increasing pixel depth can improve measurement accuracy, but it also reduces system speed and increases the difficulty of system integration (more cables, larger size, etc.).
2.4 Frame Rate
The speed at which a camera acquires and transmits images is typically measured in frames per second (Frames/Sec) for area scan cameras and in lines per second (Hz) for line scan cameras. The choice of frame rate needs to take into account the dynamic scenes being captured.
2.5 exposures
Industrial line scan cameras use line-by-line exposure, and can choose between fixed line frequency and external trigger synchronization. The exposure time can be consistent with the line period or a fixed time can be set. Area scan cameras have several common modes, including frame exposure, field exposure, and rolling exposure. Industrial digital cameras generally provide external trigger image acquisition, and the shutter speed can generally reach 10ms, with high-speed cameras being even faster.
2.6 Noise
Noise refers to unwanted signals outside the actual image target that are not captured during the imaging process. It is broadly divided into two categories: shot noise, which is introduced by the effective signal and exists in all cameras; and inherent noise unrelated to the signal, which is generated by the image sensor readout circuitry, camera signal processing, and amplification circuitry. The inherent noise varies from camera to camera.
2.7 Development Interface
For development projects that utilize cameras for visual purposes, camera control (taking photos, recording videos, saving, setting parameters, etc.) is essential. Manufacturers typically provide control demos, and during development, these control functions need to be applied to our projects.
3. Lens
Cameras and lenses are usually sold as a set. Lens selection primarily considers the image viewing distance. The lens selection steps are as follows:
1) Calculate the number of pixels corresponding to the shorter side, E=B/C. The number of pixels on both the longer and shorter sides of the camera must be greater than E;
2) Pixel size = Product short side size B / Number of pixels on the short side of the selected camera
3) Magnification = Size of selected camera chip / Field of view of the short side of the camera
4) Resolvable product accuracy = pixel size / magnification (to determine if it is less than C)
5) Objective lens focal length = working distance / (1 + 1 / magnification) Unit: mm
6) The resolution of the image plane must be greater than 1/(2 × 0.1 × magnification). Unit: lp/mm
The selected lens must support a CCD size equal to or larger than the camera's CCD sensor chip. Additionally, the mounting mount must be compatible (C, CS, or F type). Consider the lens's working distance and whether there is sufficient space. If you are still unsure about lens selection, consult the manufacturer's technical support for recommendations based on your application scenario.
4. Light source
In machine vision, supplemental lighting is an essential operation. Simply increasing the camera's exposure time will increase image noise and reduce image quality; therefore, it is necessary to select a suitable light source. The selection of a light source consists of two parts: the light source lamp and the light source controller.
4.1 Light source lamp
Machine vision offers a wide variety of light sources because its applications are very broad, and the appropriate light source must be selected based on the specific project to achieve the desired results.
The following are common types of light sources; you can choose according to the specific project.
4.2 Light Source Controller
The light source controller provides illumination for the testing environment. It typically provides a development interface to control the designated output ports to light up and turn off, mainly working with the camera to provide the software with images that meet the requirements.
5. Physical environment
Machine vision systems have high requirements for the testing environment, involving issues such as the camera, light source, and placement of the target under test. The principle is to provide a stable, high-quality image. The testing environment may require design by a professional organization; however, for simple experimental environments, experimental stands can be purchased online.
Article Source: Positive Sunshine i, Tech Success Training Network
