In machine vision, resolution is well known as an important parameter for measuring lenses and cameras. However, in actual combination applications, do you know how to effectively match lens resolution and camera resolution? First, let’s understand the concept of lens resolution

2025/10/1415:51:36 beauty 1725

In machine vision, resolution is well known as an important parameter for measuring lenses and cameras. However, in actual combination applications, do you know how to effectively match lens resolution and camera resolution?

First, let’s understand the concepts of lens resolution and camera resolution.

Lens resolution

Lens resolution refers to its ability to distinguish two close objects. In the actual application of , we generally use the minimum distance at which the "image" of a "point object" can be distinguished after being imaged by the lens. If it is less than this distance, we think it is one point; if it is greater than this distance, we think it is two separate points. In addition, there is another expression which is the number of line pairs per millimeter LP/mm. The reason why the

lens has resolution is due to aberration and diffraction (not explained in detail here), which causes distortion from "object" to "image". After the point object is imaged by the lens, it is no longer a point, but becomes a spot.

camera resolution

camera resolution refers to how many pixels are used to display an image per unit distance.

Assume that the size of pixel is 2.2μm, with a 0.5x lens. When measuring a 22μm object, since the 22μm object becomes an 11μm image after passing through the 0.5x lens, 11 μm / 2.2 μm = 5 pixels are used to display. Therefore, the image per unit distance is displayed with 11÷2.2/11=1/2.2 pixels. That is, the camera image resolution is 1/2.2 pixel/μm (actually the reciprocal of the pixel size). From this derivation we conclude that the smaller the pixel size, the higher the resolution of the camera.

After understanding the respective concepts, how do we match the lens resolution and the camera resolution?

Are there any convenient and effective matching methods?

Effective allocation of lens resolution and camera resolution

Method 1

The resolution marked in our product parameter table is the object-space resolution of the lens; multiply the object-space resolution by the magnification to get the image-space resolution; compare the calculated image-space resolution with 2 times the pixel size:

) If the pixel size of the lens is 2 times the resolution of the image side, the resolution of the camera is wasted.

Method 2

The optical image formed by the lens is itself an analog signal with infinite points. However, during the reception process of the image receiver (CCD or CMOS), discrete sampling is formed due to the spacing between pixels. The sampling law, also called Shannon's sampling law, tells us that for an analog signal with frequency f, in order to restore it without distortion, we must sample at least a frequency of 2f.

In the previous part of this article, we introduced: Another representation of lens resolution is spatial frequency, which is the number of line pairs per millimeter LP/mm. Therefore, assuming that the resolution of the lens is n LP/mm, then we must ensure that there are 2n pixels per millimeter. Only in this way can the resolving power of the lens be fully utilized.

For example: Molite ML-MC-XR series and ML-M-UR series are marked with a resolution of 200LP/mm, so a chip with 400 pixels per millimeter is used to receive images, so as not to waste the resolution of the lens. With 400 pixels per millimeter, the calculated pixel size is almost 2.2 μm, which is why we also mark that this lens supports 2.2 μm pixels.

Tips:

Lens resolution is just a parameter of the lens itself and has nothing to do with the camera.
The resolution of the camera is only a parameter of the camera itself. It is related to the pixel size and has nothing to do with the lens. The two independent resolutions of
must match so that the performance of one party will not be wasted.
Whether it is a lens or a camera, resolution is only one of the parameters for evaluating imaging quality, and the higher the resolution, the better. If you want to evaluate the overall performance, you must also consider other parameters and specific applications.

Momentview intelligent machine vision low-code platform is a cloud collaborative development platform for machine vision applications. It always adheres to the product concept of 0 cost, 0 code, 0 threshold, and 0 hardware. The

platform takes artificial intelligence technology as its core and provides developers with a one-stop complete tool chain for image collection, image annotation, algorithm development, algorithm packaging and application integration in the machine vision application development process. Covering hundreds of general functions such as character recognition, defect detection, target positioning, size measurement, video streaming , etc., it is committed to becoming the machine vision low-code platform with the largest number of users and the most extensive implementation scenarios in the world.

If you need to use similar machine vision in your industrial production line: visual inspection, visual recognition, visual positioning and other technologies, then you may wish to chat with us. We will first analyze your needs and give you a suitable solution from a professional perspective!

beauty

One day in April 1929, the editor of a Moscow newspaper was handing out assignments when he noticed that one of his reporters had not taken any notes. Then, before the editors could argue, Seleshevsky began to recite the entire meeting, down to the last detail.

Is photographic memory really a thing?

10/27 1537

As we enter 2023, the Spring Festival is not far away, and spring will soon come again. Many people have travel plans in 2023, and we can finally go where we want to go.

In 2023, you can finally go where you want to go. Do you love taking photos more?

10/27 1026

Readers, holding this first magazine of 2023 in your hands, watching the snow-capped peaks of the plateau holding up a bright moon rising, you may be filled with emotions. When we look through the photos taken by photographers, we suddenly realize that the past year recorded in t

Quick overview of new issues | "Chinese Photographers" Issue 1, 2023

10/27 1939

As long as the Qing Dynasty is mentioned, especially the late Qing Dynasty, people will always label it as backward and ignorant. Western photography technology has been introduced to China during the late Qing Dynasty. A hundred years later, we can understand the current social

Colored old photos from the late Qing Dynasty: A rich man sitting on a kang counting coins, and Kong Lingyi, the Duke of Yan, drinking with foreigners.

10/27 1919