First, visual system principle description

Machine vision is to use machines instead of human eyes to measure and judge. Machine vision system refers to the machine vision products (that is, image uptake device, divided into CMOS and CCD two) will be captured target into image signal, transmitted to a special image processing system, according to the pixel distribution and brightness, color and other information, into a digital signal; The image system performs various operations on these signals to extract the features of the target, and then controls the action of the equipment in the field according to the discriminant results.

Two, visual system components

The visual system consists mainly of the following parts

1) Lighting source

2) the camera

3) Industrial cameras

4) Image acquisition/processing card

5) Image processing system

6) Other external equipment

Three, the camera

1. Detailed introduction

Industrial camera is also known as the camera, compared with the traditional civil camera (camera), it has high image stability, high transmission ability and high anti-interference ability, etc. At present, most of the industrial cameras on the market are based on CCD (Charge Coupled Device) or CMOS (Complementary Metal Oxide Semiconductor) chips.

CCD is the most commonly used image sensor in machine vision. It integrates photoelectric conversion, charge storage, charge transfer and signal reading, and is a typical solid-state imaging device. The outstanding characteristic of CCD is to take charge as a signal, and different from other devices is to take current or voltage as a signal. This kind of imaging device forms a charge packet through photoelectric conversion, and then transfers and amplifies the output image signal under the action of driving pulse. A typical CCD camera consists of optical lens, timing and synchronization signal generator, vertical driver, analog/digital signal processing circuit. As a functional device, CCD has the advantages of no burn, no lag, low voltage operation and low power consumption compared with vacuum tube.

The development of CMOS image sensor first appeared in the early 1970s. In the early 1990s, with the development of very large scale integrated circuit (VLSI) manufacturing technology, CMOS image sensor has been rapidly developed. CMOS image sensor integrates photosensitive element array, image signal amplifier, signal reading circuit, analog-to-digital conversion circuit, image signal processor and controller on a chip, and has the advantage of programming random access to local pixels. At present, CMOS image sensors are widely used in high resolution and high speed applications due to their good integration, low power consumption, high speed transmission and wide dynamic range.

2. classification

Anything classification must have its own classification standards, industrial cameras are no exception, according to the type of chip can be divided into CCD camera, CMOS camera; According to the structural characteristics of the sensor, it can be divided into linear array camera and planar array camera. According to the scanning mode can be divided into interlaced scanning camera, line by line scanning camera; According to the resolution size can be divided into ordinary resolution camera, high resolution camera; According to the output signal mode can be divided into analog camera, digital camera; According to the output color can be divided into monochrome (black and white) camera, color camera; According to the output signal speed can be divided into ordinary speed camera, high speed camera; According to the response frequency range can be divided into visible light (ordinary) camera, infrared camera, ultraviolet camera, etc.

3. the difference between

1) The performance of industrial camera is stable and reliable and easy to install, the camera structure is compact and strong, not easy to damage, continuous working time is long, can be used in poor environment, the general digital camera is not able to do these. For example, it would be unbearable to keep a civilian digital camera working 24 hours a day or several days in a row.

2) Industrial cameras have a very short shutter time and can capture objects moving at high speed.

For example, stick a business card on the blade of an electric fan, rotate it at maximum speed, set the right shutter time, snap an image with an industrial camera, and still be able to make out the font clearly. It's impossible to get the same effect with a normal camera.

3) The image sensor of industrial camera scans line by line, while the image sensor of ordinary camera scans line by line. The production process of the image sensor of line by line scanning is complicated, with low yield and small shipment volume. Only a few companies in the world can provide this kind of product, such as Dalsa and Sony, and the price is expensive.

4) The frame rate of industrial cameras is much higher than that of ordinary cameras.

Industrial cameras can take anywhere from ten to a few hundred images per second, while ordinary cameras can only take two or three images, which is a big difference.

5) The output of industrial cameras is raw data, and its spectral range is often wide, which is more suitable for high-quality image processing algorithms, such as Machine Vision applications. However, the spectral range of pictures taken by ordinary cameras is only suitable for human vision, and after MJPEG compression, the image quality is poor, which is not conducive to analysis and processing.

6) Industrial Camera (Industrial Camera) is more expensive than ordinary Camera (DSC).

4. How to choose

1) CCD or CMOS cameras are selected according to different applications

CCD industrial camera is mainly used in the image extraction of moving objects, such as the mounter machine vision, of course, with the development of CMOS technology, many mounters are also in the selection of CMOS industrial camera. Used in the visual automatic inspection scheme or industry in general with CCD industrial camera more. CMOS industrial cameras are more and more widely used due to their low cost and low power consumption.

2) Selection of resolution

Firstly, the precision of the object to be observed or measured is considered, and the resolution is selected according to the precision. Camera pixel accuracy = single-direction field of view/single-direction camera resolution. Then the single-direction resolution of the camera is equal to the single-direction visual field/theoretical accuracy.

If the single field of view is 5mm long and the theoretical accuracy is 0.02mm, the single-direction resolution =5/0.02=250. However, in order to increase the stability of the system, it is not necessary to use one pixel unit to correspond to one measurement/observation accuracy value. Generally, multiples of 4 or higher can be selected. In this way, the camera needs a single-direction resolution of 1000, and 1.3 megapixels is enough.

Second look at the output of the industrial camera, if the type of observation or machine software analysis identification, high resolution is helpful; If VGA output or USB output, observe on the display, it also depends on the resolution of the display, industrial camera resolution again high, display resolution is not enough, is meaningless; The high resolution of industrial cameras also helps, using memory cards or photo functions.

3) Match with the lens

The sensor chip size should be less than or equal to the lens size, and the C or CS mounting mount should also match (or add a transfer port);

4) Selection of camera frame number

When the object under test has motion requirements, industrial cameras with high frame number should be selected. But generally the higher the resolution, the lower the number of frames.

Four, lens article

The basic function of the lens is to realize the beam transformation (modulation). In the machine vision system, the main role of the lens is to image the target on the photosensitive surface of the image sensor. The quality of the lens directly affects the overall performance of the machine vision system, reasonable selection and installation of the lens, is an important part of the machine vision system design.

Basic knowledge

1. Lens matching

How do you choose the right lens? When selecting the lens, you need to choose the lens that matches the size of the camera interface and CCD. The interface mode of lens C and CS is dominant. Small CS-interface cameras for security are gaining popularity, while the FA industry is mostly a combination of C-interface cameras and lenses. The corresponding CCD size, the market generally according to the use of 2/3 inch to 1/3 inch products.

2. Interchangeability

The C interface lens can be interacted with the C interface camera and the CS interface camera.

The CS interface lens cannot be applied to the C interface camera. Only the CS interface camera can be applied.

3. KERARE

If the camera uses a lens with a small CCD size, the part of the periphery that does not take in the image will appear black, we call it KERARE.

4. The role of the lens

The lens is designed by grinding various nitrate-materials with different refractive index into high-precision curved surfaces and combining these lenses. The basic principle is a common technique used since the time of Galileo. In order to get clearer images, new glass and aspheric lenses have been developed and manufactured.

Five. light source

LED light source, halogen lamp (fiber light source), high frequency fluorescent lamp. At present, LED light source is the most commonly used, which mainly has the following characteristics:

· It can be made into various shapes, sizes and irradiation angles;

· It can be made into various colors according to needs, and the brightness can be adjusted at any time;

· Through the heat dissipation device, the heat dissipation effect is better and the brightness is more stable;

· Long service life;

· Quick response, reaching maximum brightness in 10 microseconds or less;

· Power supply with external trigger, can be controlled by computer, fast starting speed, can be used as a strobe light;

· LED with low operating cost and long life will show greater advantages in terms of comprehensive cost and performance;

· Special design can be carried out according to customers' needs.

LED light sources are classified by shape

1. Ring light source

The ring light source provides different illumination angles and different color combinations, which can better highlight the three-dimensional information of the object. High density LED array, high brightness; A variety of compact design, save installation space; Solve the problem of diagonally illuminated shadow; Optional diffuse plate light guide, uniform light diffusion. Application :PCB substrate testing, IC component testing, microscope lighting, liquid crystal correction, plastic container testing, etc.

2. Back light source

High density LED array surface to provide high intensity backlighting, can highlight the object. It is especially suitable for use as a microscope carrier. Red and white back light source, red and blue multi-purpose back light source, can be configured with different colors, to meet the multi-color requirements of different measured objects. Application: MEASUREMENT OF MECHANICAL PARTS size, electronic components, IC shape detection, film stain detection, transparent object scratch detection, etc

3. Strip light source

The strip light source is the preferred light source for the object with large square structure. Color can be matched according to demand, free combination; The irradiation Angle and installation can be adjusted at will. Application: Metal surface inspection, image scanning, surface crack detection, LCD panel detection, etc.

4. Coaxial light source

The coaxial light source can eliminate the shadow caused by the uneven surface of the object, so as to reduce the interference. Part of the spectrometer design, reduce light loss, improve imaging clarity, uniform irradiation of the surface of the object.

Application: THE SERIES LIGHT SOURCE is MOST suitable FOR highly reflective objects, such as metal, glass, film, wafer surface scratch detection, chip and silicon wafer damage detection, Mark point positioning, packaging bar code recognition.

5. AOI special light

Different angles of three-color light lighting, irradiation highlights the three-dimensional information of solder; Plus diffuse plate to guide light, reduce reflection; Different Angle combination; Application: Used for solder inspection of circuit boards.

6. Spherical integral light source

The inner wall of the hemispherical surface with integral effect uniformly reflects the light emitted 360 degrees from the bottom, so that the illuminance of the whole image is very uniform. Application: Suitable for curved surface, concave and convex surface, arc surface detection, or metal, glass surface reflective object surface detection.

7. Linear light source

Ultra high brightness, using cylindrical lens to collect light, suitable for all kinds of pipeline continuous detection occasions. Application: Special for array camera lighting and AOI.

8. Point light source

High power LED, small size, high luminous intensity; Fiber halogen lamp substitute, especially suitable for lens coaxial light source; Efficient heat dissipation device, greatly improve the service life of light source. Application: Suitable for telecentric lenses, chip testing, Mark point positioning, wafer and LCD glass substrate correction.

9. Combination of strip light source

Four sides are equipped with strip light, each side of the independent and controllable lighting; The lighting Angle can be adjusted according to the requirements of the object under test, with wide applicability. Application cases :CB substrate detection, IC element detection, solder inspection, Mark point positioning, microscope lighting, packaging bar code lighting, spherical object lighting, etc.

10. Counterpoint light source

Fast counterpoint speed; The view; High precision; Small size, easy to detect and integrate; High brightness, optional auxiliary ring light source. Application :VA series light source is a special light source for full automatic circuit board printing machine.

Selection of light source

1. Prerequisite information

1) Test content

Appearance inspection, OCR, size determination, positioning

2) Objects

① What do you want to see? (foreign body, scar, defect, mark, shape, etc.)

② Surface state (mirror, rough, surface, plane)

③ 3 d? The plane?

④ Material and surface color

⑤ Range of vision?

⑥ Dynamic or static (camera shutter speed)

3) Restrictions

① Working distance (distance from the lower end of the lens to the surface of the subject)

② Setting conditions (the size of lighting, the distance between the lower end of lighting and the surface of the subject under test, reflection type OR transmission type)

③ The surrounding environment (temperature, light)

④ Type of camera, plane array or line array

2. Simple preparatory knowledge

1) Due to the different material and thickness, the light transmission characteristics (transparency) are different.

2) The length of beam and penetration ability of light root best are different after irradiation.

3) The longer the wavelength of light, the stronger the penetration of the substance, the shorter the wavelength of light and the greater the severing rate of the surface of the substance.

4) Transmission lighting, that is, the light transmission object, and tourism through the light of the lighting technique.

3. the light source

1) Stable and uniform light sources are extremely important

2) objective: will be tested and the background Ming 顕 distinguish as far as possible

3) The most important thing to allow for an image is to get a clear indication of the difference between what is being measured and the background

4) At present, the most widespread technique in the field of image processing is: binarization (white black) processing

In order to highlight the feature points, the feature image is highlighted, in the lighting technique, commonly used including bright field of vision and dark field of vision.

Clear visual field: using direct light to tourism the object as a whole (scattered light in black)

Dark vision: Using scattered light to tourism the object as a whole (direct light is white)

The specific light source selection method also depends on the practical experience of the experiment.