Photometric stereo-based single time-of-flight camera.

Area of Science:

• Computer Vision
• Optical Sensing
• Image Processing

Background:

• Time-of-flight (ToF) cameras offer depth sensing capabilities but often suffer from limited depth detail and accuracy.
• Conventional ToF systems require specialized hardware or complex post-processing to improve depth map quality.
• Enhancing depth resolution and accuracy in ToF imaging remains a significant challenge for various applications.

Purpose of the Study:

• To present a novel method for enhancing the depth quality of time-of-flight (ToF) cameras.
• To achieve simultaneous acquisition of depth and normal maps without requiring additional hardware.
• To improve the level of detail in depth maps typically unachievable with standard ToF cameras.

Main Methods:

• The proposed method involves controlling the light-emitting diode (LED) turn-off patterns of the ToF camera.
• Sixteen subphase images are captured by varying gate-pulse timing and LED emission patterns.
• Depth and normal maps are generated from these subphase images and then combined for enhanced depth detail.

Main Results:

• The integration of normal maps with depth maps significantly improves the captured depth details.
• The average absolute difference between measured and laser-scanned depth maps was reduced from 4.57 mm to 3.77 mm.
• The method successfully enhances depth quality without necessitating any additional devices or hardware modifications.

Conclusions:

• The developed technique effectively enhances the depth quality of ToF cameras through intelligent control of LED patterns.
• Simultaneous depth and normal map acquisition provides richer scene information, leading to improved 3D reconstruction.
• This hardware-free approach offers a practical solution for improving ToF camera performance in depth-sensitive applications.