Object detection

💡 Reminder: All code and scripts for this project are executed in the cloud. Open the Google Colab Notebook and select Runtime > Run all to execute the setup.

What is object detection?

Object detection identifies objects in a camera image and locates each one. For every detection, the model returns:

• A class label, such as person, ball, or helmet.
• A confidence score.
• A bounding box around the object.

Object detection differs from image classification. Classification assigns one label to an entire image. Object detection can find several objects in a single image.

How it works with Petoi AI Vision

Petoi AI Vision combines a camera with an embedded AI processor. The module captures images, runs the deployed model locally, and sends detection results to the robot controller.

The robot can use these results to react to its surroundings. For example, it can follow a ball, recognize a person, or respond to a gesture.

The final model must be optimized for the Ethos-U55 accelerator on Petoi AI Vision.