A Method to Compensate for the Errors Caused by Temperature in Structured-Light 3D Cameras.

Area of Science:

• Computer Vision
• Robotics
• Sensor Technology

Background:

• Industrial applications demand high accuracy from red-green-blue-depth (RGB-D) cameras.
• Standard calibration methods often neglect temperature's influence on camera measurements.
• Orbbec Astra cameras, while low-cost, require robust calibration for precise data.

Purpose of the Study:

• To evaluate the effect of temperature on Orbbec Astra camera measurements.
• To develop and validate a method for correcting temperature-induced errors in RGB-D cameras.
• To assess the proposed correction method's performance and applicability in industrial settings.

Main Methods:

• An experimental study was conducted in a thermal chamber to analyze camera performance under varying temperatures.
• Temperature-induced errors were modeled using a hyperbolic paraboloid function.
• A two-step compensation method was proposed, utilizing either infrared distortion or depth maps to compute and correct errors.

Main Results:

• Temperature variations were found to introduce errors that can be mathematically modeled.
• The proposed two-step correction method effectively compensated for temperature-induced errors.
• The method demonstrated good performance across multiple Orbbec Astra cameras in experimental and real-world scenarios.
• Comparison with Kinect v1 showed similar performance, indicating the method's robustness.

Conclusions:

• The developed method accurately corrects temperature-induced errors in Orbbec Astra cameras.
• The correction technique is simple, computationally inexpensive, and shows potential for application to other similar RGB-D cameras.
• This research enhances the reliability of low-cost RGB-D cameras for demanding industrial applications.