Media.Player

Visual Servo Controller to Keep Target Objects in Video Camera's Field of View

IP Title

Systems And Methods For Maintaining Multiple Objects Within A Camera Field-of-View

Detailed Technology Description

None

Application Date

Feb 1, 2011

Application No.

9,288,449

Others

*Abstract

Dynamically Controls Camera Settings to Track Multiple Objects in Motion

This visual servo is a vision-based controller that moves a video camera to keep moving target objects in its field of view by using image data and physical constraints imposed on the camera’s settings. This controller can employ pose, position and velocity to control the operation of various cameras used with drones, robots or other surveillance situations. Available visual servo systems are ineffective at maintaining multiple moving objects in the camera’s field of view because they rely on a pre-determined goal image, pose or trajectory. If the movement patterns of the objects are not plotted at fixed or rigid target points beforehand, the machine cannot track them. Multiple moving objects would not have a fixed set of points.

Researchers at the University of Florida have developed a visual servo controller that can identify and track multiple moving objects, keeping them within the camera’s field of view. This visual servo could monitor multiple people or vehicles in surveillance situations or multiple players moving in many directions in sports broadcasting. Individuals could use this visual servo at home with a recreational drone to track and record group activities.

Application

Visual servo controller that keeps multiple moving targets within a video camera’s field of view.

Advantages

• Analyzes visual feedback of objects in motion to dynamically adjust camera configurations, ensuring that the targets remain within the camera’s field of view
• Controls camera settings to regulate vision-based mechanical tasks, such as surveillance monitoring and vision tracking for mobile robots, unmanned aerial vehicles, pan-tilt-zoom cameras and more
• Computes algorithm efficiently, avoiding increased running times when adding additional target objects

Technology

Unlike most visual servo controllers, this vision-based tracking system does not use pre-determined goal images, poses, or trajectories. Rather, a set of undetermined task functions use information from visual feedback data to regulate various constraints. By prioritizing task functions, the servo regulates the mean, variance and velocity of coordinates, keeping the targeted objects centered within the camera’s point of view.

*IP Issue Date

Mar 15, 2016

*IP Publication Date

Jun 2, 2011

*Principal Investigator

Name: Warren Dixon

Department:

Name: Nicholas Gans

Department:

Country/Region

USA