Machine vision is the use of machines to replace human eyes for measurement and judgment. A machine vision system refers to the process of using machine vision products (i.e., image acquisition devices, which are divided into CMOS and CCD) to capture and transmit the target, convert it into an image signal, and then feed it to a dedicated image processing system. Based on pixel distribution and information such as brightness and color, the signal is converted into a digital signal. The image system performs various operations on these signals to extract the target's features, and then controls the operation of equipment on site based on the judgment results.
Components of the vision system:
1. Light source
2. Lens
3. Industrial cameras
4. Image acquisition/processing card
5. Image Processing System
6. Other external devices
Camera section
Industrial cameras, also known as video cameras, have advantages over traditional consumer cameras (video cameras) such as high image stability, high transmission capability, and high anti-interference capability. Currently, most industrial cameras on the market are based on CCD (Charge Coupled Device) or CMOS (Complementary Metal Oxide Semiconductor) chips.
Among them, CCD is currently the most commonly used image sensor in machine vision. It integrates photoelectric conversion, charge storage, charge transfer, and signal readout, and is a typical solid-state imaging device.
A key feature of CCDs is that they use electrical charge as a signal, unlike other devices that use current or voltage. These imaging devices generate charge packets through photoelectric conversion, which are then transferred, amplified, and output as image signals under the action of driving pulses.
A typical CCD camera consists of an optical lens, a timing and synchronization signal generator, a vertical driver, and analog/digital signal processing circuitry. As a functional device, CCDs offer advantages over vacuum tubes, including no burn-in, no hysteresis, low-voltage operation, and low power consumption.
The development of CMOS image sensors first appeared in the early 1970s. In the early 1990s, with the development of very large-scale integrated circuit (VLSI) manufacturing technology, CMOS image sensors developed rapidly.
CMOS image sensors integrate a photosensitive element array, image signal amplifier, signal readout circuit, analog-to-digital converter circuit, image signal processor, and controller onto a single chip, and also have the advantage of programmable random access to local pixels.
Currently, CMOS image sensors are widely used in high-resolution and high-speed applications due to their excellent integration, low power consumption, high-speed transmission, and wide dynamic range.
Classification:
Everything has its own classification criteria, and industrial cameras are no exception.
Based on chip type, cameras can be divided into CCD cameras and CMOS cameras;
Based on the structural characteristics of the sensor, cameras can be divided into linear scan cameras and area scan cameras.
According to the scanning method, cameras can be divided into interlaced scanning cameras and progressive scanning cameras;
Based on resolution, cameras can be divided into standard resolution cameras and high resolution cameras.
According to the method of output signal, cameras can be divided into analog cameras and digital cameras;
Based on the output color, cameras can be divided into monochrome (black and white) cameras and color cameras;
Based on the speed of the output signal, cameras can be divided into ordinary speed cameras and high-speed cameras.
Based on their response frequency range, cameras can be classified into visible light (ordinary) cameras, infrared cameras, ultraviolet cameras, etc.
the difference:
1. Industrial cameras offer stable and reliable performance, are easy to install, have a compact and robust structure that is not easily damaged, and can operate continuously for long periods, even in harsh environments—features that ordinary digital cameras cannot match. For example, a consumer digital camera would certainly not be able to withstand working 24 hours a day or for several consecutive days.
2. Industrial cameras have extremely short shutter speeds, allowing them to capture fast-moving objects. For example, attach a business card to the blades of a fan and spin it at maximum speed. With an appropriate shutter speed, an industrial camera can capture an image that still clearly shows the text on the card. This effect cannot be achieved with a regular camera.
3. Industrial camera image sensors use progressive scan, while ordinary camera image sensors use interlaced scan. Progressive scan image sensors have a more complex manufacturing process, lower yield, and smaller shipment volume. Only a few companies in the world can provide such products, such as Dalsa and Sony, and they are expensive.
4. Industrial cameras have a much higher frame rate than ordinary cameras. Industrial cameras can capture ten to several hundred images per second, while ordinary cameras can only capture 2-3 images, a significant difference.
5. Industrial cameras output raw data, which often has a wide spectral range, making it suitable for high-quality image processing algorithms, such as machine vision applications. In contrast, images taken by ordinary cameras have a spectral range suitable only for human vision and are compressed using MJPEG, resulting in poor image quality and making them unsuitable for analysis and processing.
6. Industrial cameras are more expensive than regular cameras (DSC).
How to choose:
1. Choose between CCD or CMOS cameras depending on the application. CCD industrial cameras are mainly used for image extraction of moving objects, such as in pick-and-place machines. However, with the development of CMOS technology, many pick-and-place machines are also choosing CMOS industrial cameras. CCD industrial cameras are generally more commonly used in visual automated inspection solutions or industries. CMOS industrial cameras are becoming increasingly widely used due to their lower cost and lower power consumption.
2. Resolution Selection: First, consider the precision of the object to be observed or measured, and select the resolution accordingly. Camera pixel precision = Single-direction field of view size / Camera single-direction resolution. Therefore, camera single-direction resolution = Single-direction field of view size / Theoretical precision. If the single field of view is 5mm long and the theoretical precision is 0.02mm , then the single-direction resolution = 5/ 0.02 = 250. However, to increase system stability, a single pixel unit is not used to correspond to a single measurement/observation precision value; generally, a multiplier of 4 or higher is chosen. Thus, this camera requires a single-direction resolution of 1000, and 1.3 megapixels is sufficient.
Secondly, consider the output of the industrial camera. For stereoscopic observation or machine software analysis and recognition, high resolution is helpful. For VGA or USB output, which is viewed on a monitor, the resolution of the monitor is also crucial. Even if the industrial camera has a high resolution, it is meaningless if the monitor resolution is insufficient. When using memory cards or taking photos, a high resolution of the industrial camera is also helpful.
3. The sensor chip size must be smaller than or equal to the lens size, and the C or CS mounting bracket must also be compatible (or an adapter can be added).
4. Camera Frame Rate Selection: When the object being measured requires motion, an industrial camera with a high frame rate should be selected. However, generally speaking, the higher the resolution, the lower the frame rate.
Lens Section
The basic function of a lens is to achieve beam transformation (modulation). In machine vision systems, the main role of the lens is to project the image target onto the photosensitive surface of the image sensor. The quality of the lens directly affects the overall performance of the machine vision system; therefore, the proper selection and installation of the lens is a crucial aspect of machine vision system design.
Basic knowledge:
Lens Matching
When choosing a lens, it's crucial to select one that matches the camera's interface and CCD size. C-mount and CS-mount lenses are the most common interface types. Small security cameras with CS-mount interfaces are widely used, while the FA (Film/Analog Device) industry primarily uses C-mount cameras and lenses. Regarding the corresponding CCD size, the market generally uses 2/3-inch to 1/3-inch products depending on the application.
Interchangeability
C-mount lenses are compatible with both C-mount and CS-mount cameras.
CS interface lenses cannot be used with C interface cameras; they can only be used with CS interface cameras.
KERARE
If a camera uses a lens with a small CCD size, the areas around the camera that are not captured will appear black, which we call KERARE.
The function of the lens:
Lens design involves grinding various nitrate materials with different refractive indices into high-precision curved surfaces and then assembling these lenses. The basic principle is based on a technique widely used since the time of Galileo. To obtain sharper images, research, development, and trial production of new nitrate materials and aspherical lenses have been ongoing.
Light source section
LED light sources, halogen lamps (fiber optic lamps), and high-frequency fluorescent lamps. Currently, LED light sources are the most commonly used, and they have the following main characteristics:
• Can be made into various shapes, sizes and various irradiation angles;
• It can be made in various colors as needed, and the brightness can be adjusted at any time;
• The heat dissipation device improves heat dissipation and makes the light brightness more stable;
Long service life;
• It has a fast response time and can reach maximum brightness in 10 microseconds or less;
• The power supply has an external trigger, can be controlled by a computer, has a fast start-up speed, and can be used as a strobe light;
LEDs with low operating costs and long lifespan will have a greater advantage in terms of overall cost and performance;
Custom designs are available to meet specific customer needs.
LED light sources can generally be classified into the following categories according to their shape:
1. Ring Light Source: Ring light sources offer different illumination angles and color combinations, better highlighting the three-dimensional information of objects; high-density LED array, high brightness; various compact designs save installation space; solves the problem of shadows from diagonal illumination; optional diffuser plate for uniform light diffusion. Applications : PCB substrate inspection, IC component inspection, microscope illumination, LCD calibration, plastic container inspection, integrated circuit printing inspection.
2. The backlight uses a high-density LED array to provide high-intensity backlight illumination, highlighting the shape and contour features of objects, making it particularly suitable as a microscope stage. Red-white dual-use backlights and red-blue multi-use backlights can be mixed to produce different colors to meet the multi-color requirements of various measured objects. Applications include : measurement of mechanical parts dimensions, shape inspection of electronic components and ICs, film stain detection, and scratch detection on transparent objects.
3. Bar Light Source: Bar light sources are the preferred light source for larger square-shaped objects being measured; colors can be customized and freely combined; the illumination angle and installation are adjustable. Applications : metal surface inspection, image scanning, surface crack detection, LCD panel inspection, etc.
4. Coaxial Light Source: Coaxial light sources can eliminate shadows caused by uneven object surfaces, thus reducing interference; some employ a beam splitter design to reduce light loss, improve imaging clarity, and uniformly illuminate the object surface. Applications : This series of light sources is best suited for highly reflective objects, such as scratch detection on metal, glass, film, and wafer surfaces; damage detection of chips and silicon wafers; mark point positioning; and packaging barcode recognition.
5. AOI-specific light source with tri-color illumination at different angles to highlight the three-dimensional information of the solder; additional diffuser plate to guide light and reduce reflection; different angle combinations; application areas : used for solder inspection of circuit boards.
6. The spherical integrating light source has a hemispherical inner wall with an integrating effect, uniformly reflecting light emitted from the bottom 360 degrees, making the illumination of the entire image very uniform. Applications : Suitable for inspecting curved surfaces, uneven surfaces, arc-shaped surfaces, or highly reflective surfaces such as metals and glass.
7. The linear light source boasts ultra-high brightness and utilizes a cylindrical lens for focusing, making it suitable for various continuous inspection applications on production lines. Applications include : dedicated illumination for array cameras and AOI (Automated Optical Inspection) systems.
8. High-power LED point light source: small size, high luminous intensity; a replacement for fiber optic halogen lamps, especially suitable as a coaxial light source for lenses; efficient heat dissipation device greatly improves the lifespan of the light source. Applications : Suitable for telecentric lenses, used for chip inspection, mark point positioning, and wafer and LCD glass substrate calibration.
9. The combined bar light source features bar lights on all four sides, with independent and controllable illumination on each side; the required illumination angle can be adjusted according to the requirements of the object being measured, making it widely applicable. Applications include : CB substrate inspection, IC component inspection, solder inspection, mark point positioning, microscope illumination, packaging barcode illumination, spherical object illumination, etc.
10. Alignment Light Source: Fast alignment speed; large field of view; high precision; small size, easy for inspection and integration; high brightness, with optional auxiliary ring light source. Application Areas : The VA series light source is a dedicated light source for alignment in fully automatic circuit board printing machines.
Light source selection
I. Prerequisite Information
1. Testing Content
Visual inspection, OCR, dimensional measurement, positioning
2. Object
① What do you want to see? (Foreign objects, scratches, defects, markings, shapes, etc.)
② Surface condition (mirror finish, rough surface, curved surface, flat surface)
③ Three-dimensional? Two-dimensional?
④ Material and surface color
⑤ Field of view?
⑥ Dynamic or static (camera shutter speed)
3. Restrictions
① Working distance (distance from the bottom of the lens to the surface of the object being measured)
② Setting conditions (size of the illumination, distance from the lower end of the illumination to the surface of the object being measured, reflective or transmissive type)
③ Surrounding environment (temperature, external light)
④ Camera type: area scan or linear scan
II. Basic Preliminary Knowledge:
1. Due to differences in material and thickness, their light transmission characteristics (transparency) vary.
2. The wavelength of light varies, as does its ability to penetrate matter (transmittance).
3. The longer the wavelength of light, the stronger its penetrating power through matter; the shorter the wavelength of light, the greater its scattering rate on the surface of matter.
4. Transmitted illumination is a lighting technique that allows light to pass through an object and then observes the transmitted light.
III. Light source:
1. A stable and uniform light source is extremely important.
2. Purpose: To distinguish the object being measured from the background as clearly as possible.
3. When capturing an image, the most important aspect is how to clearly obtain the contrast between the density of the object being measured and the background.
4. Currently, the most widely used technique in the field of image processing is binarization (white-black) processing. In order to highlight feature points and make the feature image stand out, the commonly used lighting techniques include bright field and dark field.
Clear vision: using direct light to observe the entire object (scattered light appears black).
Dark field of view: using scattered light to observe the object as a whole (direct light appears white). The specific method for selecting the light source depends on practical experience in experimentation.
Disclaimer: This article is a reprint. If it involves copyright issues, please contact us promptly for deletion (QQ: 2737591964). We apologize for any inconvenience.
