How many of these 31 questions do you know about industrial cameras?

Some machine vision engineers often think that industrial cameras with USB interfaces will cause frame loss. Generally speaking, industrial camera frame loss has nothing to do with the transmission interface used by the industrial camera, whether it is USB, 1394, GigE, or CameraLink. Poorly designed drivers or industrial camera hardware are the real cause of frame loss: The reason why poorly designed industrial cameras experience frame loss is actually because the data channel is blocked and cannot be processed in a timely manner. Therefore, when a new image comes in, the previous one may be forced to be discarded, or the new image may be forced to be discarded. To solve this problem, designers need to carry out precise design for each link of data transmission between the driver and industrial camera hardware.

In machine vision inspection technology, the input and output interfaces of industrial cameras include Camera Link, IEEE 1394, USB2.0, Ethernet, and USB3.0;

The first step is to choose the right lens. Lens selection should follow the following principles:

1. What is the chip size of the matching camera?
2. What type of interface does the camera have? C interface, CS interface or other interfaces;
3. The working distance of the lens;
4. Lens field of view;
5. Lens spectral characteristics;
6. Lens distortion rate;
7. Lens mechanical structure dimensions;

When choosing a CCD camera, you should consider the following aspects:

1. Photosensitive chip type; CCD or CMOS
2. Video characteristics; including click frequency and line frequency.
3. Signal output interface;
4. Camera working modes: continuous, trigger, control, asynchronous reset, long-time integration.
5. Video parameter adjustment and control methods: Manual, RS232.

At the same time, when choosing a CCD, you should pay attention to l inch=16mm instead of 25.4mm.

1. Imaging process

The principle of photoelectric conversion of CCD and CMOS image sensors is the same. The main difference between them is the different signal readout process. Since CCD has only one (or a few) output nodes for unified readout, the consistency of its signal output is very good. In the CMOS chip, each pixel has its own signal amplifier, which performs charge-to-voltage conversion, and the consistency of its signal output is poor. However, in order to read out the entire image signal, CCD requires a wider signal bandwidth of the output amplifier. In a CMOS chip, the bandwidth requirement of the amplifier in each pixel is lower, which greatly reduces the power consumption of the chip. This is the main reason why the power consumption of CMOS chips is lower than that of CCD. Although power consumption is reduced, the inconsistency of millions of amplifiers results in higher fixed noise, which is an inherent disadvantage of CMOS relative to CCD.

2. Integration

From the perspective of manufacturing process, the circuits and devices in CCD are integrated in the semiconductor single crystal material manufacturer, and the process is relatively complex. There are only a few manufacturers in the world that can produce CCD wafers, such as DALSA, SONY, Panasonic, etc. The CCD can only output analog electrical signals, which requires subsequent address decoder, analog converter, and image signal processor processing. It also needs to provide three sets of power supply synchronization clock control circuits with different voltages, and the integration level is very low. CMOS is integrated on a single material called metal oxide. This process is the same as the process used to produce tens of thousands of semiconductor integrated circuits such as computer chips and storage devices. Therefore, the cost of sound field CMOS is much lower than that of CCD. At the same time, the CMOS chip can integrate the image signal amplifier, signal reading circuit, A/D conversion circuit, image signal processor and controller into one chip. Only one chip can realize all the basic functions of the camera. The integration level is very high. This is where the concept of chip-level cameras came from. With the continuous development of CMOS imaging technology, more and more companies can provide high-quality CMOS imaging chips, including: Micron, CMOSIS, Cypress, etc.

3. Speed

CCD uses photosensitive output one by one, and can only output according to the prescribed program, and the speed is slow. CMOS has multiple charge-voltage converters and row-column switch control, and the readout speed is much faster. Currently, most high-speed cameras above 500fps are CMOS cameras. In addition, the CMOS address strobe switch can be randomly sampled to realize sub-window output, and higher speed can be achieved when only sub-window images are output.

4. Noise

CCD technology developed earlier and is relatively mature. It uses PN junction or silicon dioxide (SiO2) isolation layer to isolate noise. The imaging quality has certain advantages over CMOS photoelectric sensors. Due to the high integration level of CMOS image sensors, the distance between components and circuits is very close, the interference is serious, and the noise has a great impact on the image quality.

In recent years, with the continuous development of CMOS circuit noise reduction technology, good conditions have been provided for the production of high-density and high-quality CMOS image sensors.

• Resolution
• Speed ​​(frame rate/line rate)
• Noise
• Signal-to-noise ratio
• Dynamic range
• Pixel depth
• Spectral response
• Optical interface

Resolution is the most basic parameter of the camera. It is determined by the resolution of the chip used in the camera. It is the number of pixels arranged on the target surface of the chip. Usually the resolution of an area scan camera is expressed by two numbers: horizontal and vertical resolution, such as: 1920 (H) x 1080 (V). The first number indicates the number of pixels in each row, that is, there are 1920 pixels in total, and the following number indicates the number of rows of pixels, that is, 1080 rows. Nowadays, the resolution of a camera usually indicates how many K, such as 1K (1024), 2K (2048), 3K (4096), etc. When acquiring images, the resolution of the camera has a great impact on the image quality. When imaging the same large field of view (scenery range), the higher the resolution, the more obvious the details will be displayed.

The frame rate/line frequency of a camera indicates the frequency at which the camera collects images. Usually, area array cameras are expressed in frame rate, and the unit is fps (Frame Per second). For example, 30fps means that the camera can collect up to 30 frames of images in 1 second. Line array cameras usually use line frequency in units of KHz. For example, 12KHz means that the camera can collect up to 12,000 lines of image data in 1 second. Speed ​​is an important parameter of the camera. In practical applications, it is often necessary to image moving objects. The speed of the camera needs to meet certain requirements in order to image objects clearly and accurately. The frame rate and line rate of the camera are first affected by the frame rate and line rate of the chip. The maximum design speed of the chip is mainly determined by the highest clock that the chip can withstand.

The noise of industrial cameras refers to signals that are not expected to be collected during the imaging process and are outside the actual imaging target. According to the European camera testing standard EMVA1288, the noise in the camera can be generally divided into two categories: one is the statistical fluctuation noise caused by the effective signal that conforms to the Poisson distribution, also called shot noise. This kind of noise is the same for any camera, unavoidable, and has a certain calculation formula. (That is: the square of the noise = the mean of the signal). The second type is the inherent signal-independent noise of the camera itself. It is the noise caused by the image sensor readout circuit, camera signal processing and amplification circuit, etc. The inherent noise of each camera is different. In addition, for digital cameras, quantization noise will be generated when analog conversion of video signals. The higher the number of quantization bits, the lower the noise.

The signal-to-noise ratio of a camera is defined as the ratio of signal to noise in the image (the ratio of the average gray value of the effective signal to the root mean square of the noise), which represents the quality of the image. The higher the signal-to-noise ratio of the image, the better the image quality.

The dynamic range of the camera indicates the range in which the camera detects light signals. The dynamic range can be defined in two ways. One is the optical dynamic range, which refers to the ratio of the maximum light intensity at saturation to the light intensity equivalent to the noise output, which is determined by the characteristics of the chip. The other is electronic dynamic range, which refers to the ratio between saturation voltage and noise voltage. For a fixed camera, its dynamic range is a fixed value and does not change with changes in external conditions. In terms of linear response, the dynamic range of the camera is defined as the ratio of saturation exposure to noise equivalent exposure: dynamic range = full well capacity of the photosensitive element/equivalent noise signal. Dynamic range can be expressed in multiples, dB or Bit. With a large dynamic range, the camera has a stronger ability to adapt to different light intensities.

The digital signal output by a digital camera, that is, the pixel gray value, has a special number of bits, called pixel depth. For black and white cameras, the value range is usually 8-16bit. Pixel depth defines the number of gray levels from dark to bright. For example, for an 8-bit camera 0 represents full darkness and 255 represents full brightness. Numbers between 0 and 5 represent certain brightness indicators. 10bit data has 1024 gray levels and 12bit data has 4096 gray levels. For every application we carefully consider whether very fine grayscale levels are required. Upgrading from 8bit to 10bit or 12bit can indeed enhance the measurement accuracy, but it also reduces the speed of the system and increases the difficulty of system integration (increased cables and larger sizes), so we must choose carefully.

The interface refers to the interface between the camera and the lens. Commonly used lens interfaces include C-mount, CS-mount, and F-mount.

• Classification according to chip structure: CCD camera & CMOS camera
• According to the sensor structure: area scan camera & line scan camera
• Classification according to output mode: analog camera & digital camera
• Color camera & black and white camera

• The shutter time of industrial cameras is particularly short and can clearly capture fast-moving objects, while ordinary cameras are very blurry when capturing fast-moving objects;
• The image sensor of an industrial camera is progressive scan, while the image sensor of an ordinary camera is interlaced, or even scanned every three lines;
• The shooting speed of industrial cameras is much higher than that of ordinary cameras; industrial cameras can take ten to hundreds of pictures per second, while ordinary cameras can only take 2-3 images;
• Industrial cameras output naked data, and their spectral range is often relatively wide, which is more suitable for high-quality image processing algorithms and is commonly used in machine vision systems. As for the pictures taken by ordinary cameras, their spectral range is only suitable for human vision, and they have been compressed by MPEG, so the image quality is also poor;

1. Calculate the resolution: Divide the width by the minimum detection accuracy to get the pixels required for each line.
2. Detection accuracy: The actual detection accuracy is obtained by dividing the width by the pixels.
3. Number of scanning lines: Divide the length of movement speed per second by the accuracy to get the number of scanning lines per second.

Examples of selecting a line array camera based on the above calculation results are as follows:

For example, if the width is 1600 mm, the accuracy is 1 mm, and the movement speed is 22000mm/s, the camera: 1600/1=1600 pixels is at least 2000 pixels, and the camera is selected as 2k. 1600/2048=0.8. The actual accuracy is 22000mm/0.8mm=27.5KHz. The camera should be selected as 2048 pixels 28kHz camera.

• Line scan sensors used by line array cameras usually only have one row of photosensitive units (a few color line array sensors use three rows of photosensitive units)
• The line scan camera only collects one line of images at a time;
• The line scan camera only outputs one line of images at a time;
• Compared with traditional area array cameras, area array scanning collects several lines of images each time and outputs them in frames.

• Line scan cameras have higher resolution; each row of pixels of a line scan camera is generally 1024, 2048, 4096, 8012; while a general area scan camera is only 640, 768, 1280, and area arrays larger than 2048 are rare.
• The acquisition speed of line array cameras is faster; the acquisition speed of different models of line array cameras ranges from 5,000 lines per second to 60,000 lines per second. Users can choose a few lines or every dozen lines to form a frame of image for processing, so a very high frame rate can be achieved.
• Line array cameras can continuously collect and process without interruption; line array cameras can continuously collect linearly moving objects (linear guides, paper on rollers, fabrics, printed matter, objects on conveyor belts, etc.).
• Line scan cameras have a simpler and more reasonable structure. Compared with area scan cameras, line scan cameras do not waste resolution and collect useless data.

A smart industrial camera is not a simple camera, but a highly integrated micro machine vision system. It integrates image acquisition, processing and communication functions into a single camera, thereby providing a multi-functional, modular, highly reliable and easy-to-implement machine vision solution. Intelligent industrial cameras are generally composed of image acquisition units, image processing units, image processing software, network communication devices, etc. Due to the application of the latest DSP, FPGA and large-capacity storage technology, its level of intelligence continues to increase and can meet a variety of machine vision application needs.

The two main types of photoelectric sensing chips used in machine vision are CCD chips and CMOS chips. CCD is the abbreviation of ChargeCoupled Device and CMOS is Complementary Metal-Oxide-Semiconductor T