Distance measurement based on light field geometry and ray tracing

Area of Science:

• Optics
• Computer Vision
• Image Processing

Background:

• Light field cameras capture 4D data, enabling depth estimation.
• Accurate depth measurement is crucial for various computer vision applications.
• Existing methods for light field depth estimation have limitations in accuracy and generalizability.

Purpose of the Study:

• To propose a novel geometric optical model for measuring object plane distances in light field images.
• To enhance the accuracy and performance of depth estimation in light field imaging.
• To provide a robust model applicable across general imaging ranges.

Main Methods:

• Development of a geometric optical model based on ray tracing.
• Construction of two sub-models: object space model and image space model.
• Utilizing the refraction theorem, refocusing, and similar triangle principles for derivation.

Main Results:

• The proposed model accurately calculates the relationship between object imaging diameters and their corresponding distances.
• Experimental validation using plenoptic 1.0 cameras and a preliminary imaging system.
• Demonstrated superior accuracy and good performance compared to existing approaches.

Conclusions:

• The geometric optical model provides an accurate and effective method for depth estimation in light field images.
• The model shows significant improvements over current techniques, particularly in accuracy.
• The proposed approach is suitable for general imaging scenarios and offers enhanced performance.