120 or 60 fps webcam?

Is it possible to get 60 fps or 120 fps with webcam for fast motion? My camera is 60 fps webcam. I use cv2 to open live webcam and detect objects. But the program camera is not 60 fps. It is like 30 fps.

The frame rate (fps) you can achieve with your webcam and object detection model depends on several factors. These include the capabilities of your webcam, the efficiency of your object detection model, and the computational resources of your system (CPU and GPU).

In the forum post here, a user mentions that the fps values are dependent on the host system you are running the model on. Your CPU and GPU are major factors in the speed, as well as the type of stream you are using.

If you’re using OpenCV’s cv2.VideoCapture(0) to capture video from your webcam, it’s worth noting that the default frame rate might be lower than your webcam’s maximum fps. You might need to adjust the settings of cv2.VideoCapture to capture at higher frame rates.

However, even if your webcam can capture at 60 fps or 120 fps, achieving such high frame rates in real-time object detection can be challenging. Object detection models can be computationally intensive, and processing each frame might take longer than the time between frames at such high fps.

If you’re experiencing low fps, you might want to consider optimizing your object detection model or using a more efficient model. Roboflow provides a variety of object detection models with different trade-offs between speed and accuracy.

Remember that achieving high fps in object detection is not just about capturing video at high fps, but also processing each frame quickly enough. If the object detection takes too long, it will lower the effective fps of your application.


Hi! I’m Roboflow’s AI powered bot. This is an automated response powered using AI with knowledge from our docs, blog, and forum. If this solves your issue, please mark it as the solution using the checkbox icon. If this doesn’t solve your issue, please follow up with what you’ve tried/why, and the community will continue to chime in as usual.

Out of curiosity, what do you need 60 fps for? What deployment method are you using?

Some ways to speed your model up:

  1. Detect on smaller images.
  2. Use a lighter-weight model like our Fast object detection model or classification.

We want to classify objects that are running on the roller machine. And we used object detection method in the roboflow. We want to make not only the classification but also the detection on the roller machine. So the camera is a little close to the machine above. We tested with both video (60fps) and live webcam (30fps or below) and the video results are very good and correct like above 90%. But in webcam mode, the results are not good. Sometimes it is difficult to detect and miss the detection.

Are you using our web browser deploy method, or does webcam just refer to the format of incoming data? And is the poor performance due to frame rate, accuracy, or a combination?

Do you genuinely need 60 fps for this use case? Will you miss detections if you operate at 45 or 30 fps?

I use cv2 library in python. And yes, some of the objects are not even detection with 30 fps webcam. And I only can get 30 fps with webcam. I also try with web browser deploy method and the result is similar.