Machine vision system is widely used, for all kinds of applications also extend different needs, and camera selection has always been one of the most important precursors before the implementation of the plan.

First, a few questions. What is your application? What is under test?

The following basic conditions will help you select the camera you need more efficiently:

Color v. S. black and white?

Generally speaking, the accuracy of the black and white camera will be higher than the color camera, low noise, if your subject does not need to test the color difference, the black and white camera will be the best choice, especially the object surface detection, the image accuracy and detail taken by the black and white camera is usually much higher than the color camera.

Tip: whether through CCD or CMOS image sensor, the image is black and white. In order to present color images, the traditional way is to combine three images into a color image using three photoreceptors through a prism. However, because of the high cost of this method, most of the current methods are to insert a filter (Bayer filter) in front of the original CMOS and CCD sensors, and the filter spot on the filter corresponds to the pixel on the sensor. At the same resolution, the accuracy and accuracy of the black and white camera will be much higher, and it will be a better choice for applications that require edge detection and detail defect detection.

Sensor (CCD V.S. CMOS)

Image sensors are mainly divided into CCD and CMOS. CCD and CMOS image sensors have similar photosensitive principles, but their main differences are in the different transmission modes of digital signals. Because of the differences in structure and working principle, CCD and CMOS image sensors have different characteristics. CCD image sensor is superior to CMOS image sensor in sensitivity, resolution and noise control, while CMOS image sensor has the characteristics of low cost, low power consumption and high integration degree. In recent years, CMOS image sensor technology has gradually improved the resolution and noise control of the weak shortcomings, which can be used as the first choice of selection.

Global shutter V.S shutter?

The global shutter is exposed at the same time by all pixels in each row on the chip. All pixels of the sensor collect light at the same time and are exposed at the same time. When the preset exposure time is up, the sensor stops collecting light and converts the exposure image into an electronic one. In the rolling shutter mode, when the chip begins to be exposed, the sensor senses line by line in order, and the image output is carried out until each line is exposed from top to bottom. The exposure time of each line is different. As the exposure time of the shutter is longer than that of the global shutter, when shooting moving objects, it may elongate the image due to too fast, resulting in blurred shadows. The global shutter exposure time is shorter, which can solve the image dragging phenomenon. However, compared with the rolling shutter, the traditional global shutter has higher noise. With the technological upgrade, the image level generated by the global shutter has caught up with the rolling shutter shutter.

Camera port (GigE, USB)

At present, industrial cameras can be divided into three main types: USB Camera, GigE Camera and Camera Link. Camera Link cameras have higher specifications, but they also cost a lot more. At present, USB and GigE are still recommended for most detection. Here, USB and GigE interface are used for simple comparison. GigE camera can provide long transmission distance. And USB camera can plug and play, cost-effective, fast transmission. How to choose the right camera depends on the user's needs.

Tip: In normal specification, the effective transmission distance is up to 3 meters for USB camera, and up to 100 meters for GigE camera.

resolution ratio

The resolution requirement depends on the application scheme and the working distance, size, field of view, accuracy and so on. When selecting the desired resolution, you will first know the range of view and working distance of the subject. In some cases, the user will also first ask how much accuracy is required. Based on these conditions, the required resolution is calculated. For example, if the shooting field of view is 8*6mm, the working distance is 100mm, and the accuracy is 0.005mm, the resolution must be at least (8/0.005)*(6/0.005)=1,920,000, so at least 1.9MP camera should be selected. 60mm lens is more suitable (the following diagram and formula for reference). However, due to price factors and frame speed requirements, the higher the resolution, the lower the frame rate, so the user has to make their own decision. However, if the object under test is a precision device, it is another matter to consider the application requirements.

Frame rate (speed)

The frame rate is the number of frames (pieces) that can be captured or collected per second, which is another key point of selection. Because industrial cameras want to capture dynamic objects under test (whether in the production line or moving objects), the frame rate will be higher than consumer cameras or WebCAM. When selecting a camera at frame rate, you need to select a frame rate that is a little higher than the moving speed of the subject, but not much more. In general, the higher the resolution, the lower the frame rate.