When working on machine vision projects, many people encounter these issues:

· The camera is rated for 100 fps, but it can't actually reach that speed;

· As soon as you increase the exposure time, the frame rate immediately drops.

· After the exposure was shortened, the image became so dark that it was impossible to see clearly.

· Even though the frame rate is high, fast-moving objects still have motion blur.

The issue usually can't be avoided without addressing two parameters:

Frame rate and exposure time.

What exactly is their relationship? Let’s clarify it all in today’s article.

1. What are frame rate and exposure time, respectively?

1. Frame rate: How many frames are captured per second?

Frame rate is usually expressed in FPS.

For example, 30fps means capturing 30 frames per second.

The higher the frame rate, the smoother the footage, making it more suitable for capturing fast-moving objects, such as:

· Pipeline component inspection;

· Traffic photo capture;

· Robotic arm positioning;

· High-speed experimental record.

But the higher the frame rate, the less time each frame gets.

2. Exposure time: How long each image is “exposed to light”

The longer the exposure time, the more light enters, and the image generally becomes brighter.

But if the exposure is too long, fast-moving objects are prone to motion blur.

Simply put:

Frame rate determines how densely the footage is captured, while exposure time determines whether each individual frame comes out blurry.

This point is very important.

II. What is the relationship between frame rate and exposure time?

You can start by using a simple formula for estimation:

Exposure time (μs) ≤ 1,000,000 ÷ Frame rate (fps)

For example, the camera needs to run at 50 fps:

1,000,000 ÷ 50 = 20,000 μs

In other words, each frame lasts only 20 milliseconds.

If the exposure time is set to 50 ms, then the theoretical maximum frame rate is only:

1,000,000 ÷ 50,000 = 20 fps

Therefore, the longer the exposure time, the lower the frame rate usually is.

This is the most fundamental relationship between the two.

3. Why is the actual frame rate often lower than the calculated value?

Because every time a camera takes a picture, it’s not just about exposure.

It also needs to complete:

· Sensor readout;

· Image transmission;

· Data cache;

· Computer receives;

· Software processing.

So, what’s closer to the actual situation is:

Single-frame time = Exposure time + Readout time + Other overheads

For example, a camera can achieve up to 53 fps at full resolution.

Theoretical single-frame period is approximately:

10,000,000 ÷ 53 ≈ 18,868 μs

But this doesn't mean that the exposure time can be set directly to 18,868 μs.

Because we still need to leave time for image reading and transmission.

Therefore, when it comes to camera parameters, you can’t just rely on formulas—ultimately, you still need to take a look at:

· Product specification sheet;

· Actual capture frame rate;

· Does the software drop frames?

4. Besides exposure, what else can affect the frame rate?

1. Resolution

The larger the image, the more data there is, and the lower the frame rate.

If the project only detects local regions, you can enable ROI cropping, which often significantly improves the frame rate.

2. Image format

The data volumes for Mono8, Mono12, and RGB8 are different.

Color images typically consume more bandwidth than grayscale images.

3. Camera interface

USB3.0, GigE, Camera Link, and CoaXPress have vastly different bandwidths.

No matter how fast the sensor is, it’s useless if the interface can’t keep up.

4. Host computer performance

The CPU, memory, network card, hard disk, and algorithm speed can all potentially cause frame drops.

Sometimes it’s not that the camera is shooting too slowly—it’s that the computer can’t handle it.

5. What should I do if I want a high frame rate but the image is too dark?

Many people’s first reaction is to increase the gain.

Actually, the more reasonable order is:

1. First, strengthen the light source.

You can try:

· Increase the brightness of the light source;

· Adjust the irradiation angle;

· Shorten the distance to the light source;

· Use a more suitable ring light, strip light, backlight, or coaxial light;

· Use a strobe light source.

In high-speed scenarios, strobe lights are often very effective.

It can deliver high brightness in a very short time, effectively “freezing” fast-moving objects.

2. Adjust the lens aperture again.

When the aperture is opened wider, the amount of light entering increases.

However, an excessively large aperture can also lead to issues such as a shallower depth of field and more difficult focusing.

So it’s not always better to go bigger—instead, you need to strike a balance based on the specific site conditions.

3. Finally, add gain.

Gain can brighten the image, but it also amplifies noise at the same time.

With excessive gain, the image is prone to:

· Granular;

· Dark noise;

· Edge instability;

· The defect features are drowned out by noise.

Therefore, gain is better suited for final compensation; it’s not recommended to max it out right from the start.

Six: Does a high frame rate guarantee clear footage of fast-moving objects?

Not necessarily.

For example, if the camera is running at 100 fps, the period for each frame is 10 ms.

Even if the exposure time is also 10 ms, fast-moving objects can still exhibit motion blur.

Conversely, if the camera has only 30 fps but an exposure time of just 100 μs, a single frame could actually be remarkably sharp.

So remember:

High frame rate ensures continuity, while short exposure ensures clarity.

If you want both continuous shooting and clear images, you’ll need:

· High frame rate;

· Short exposure;

· Strong light source;

· Suitable lens;

· Sufficient transmission bandwidth.

Essential and indispensable.

7. On-site parameter tuning—recommended to follow this order.

First, shorten the exposure time based on the object’s speed to ensure there’s no motion blur.

Step 2: Set the required frame rate based on the production takt time—don't blindly aim for the highest possible value.

Step 3: Use a light source and aperture to adjust the brightness.

Step 4: If it’s still not bright enough, slightly increase the gain.

Step 5: Check the actual output frame rate, the software’s received frame rate, and the frame loss situation.

Don't just look at the numbers on the parameter panel.

The camera displays 100 fps, but that doesn't mean the algorithm is actually processing at 100 fps.

Summary

Finally, remember these four sentences:

1. Frame rate determines how many images are captured per second.

2. The exposure time determines whether a single image will be blurry.

3. The longer the exposure, the lower the frame rate usually is.

4. If you want both high speed and clarity, prioritize improving the light source—don't rely solely on gain.

Adjusting the parameters of an industrial camera essentially involves finding a balance among speed, brightness, sharpness, and noise.