Displacement equation-free laser triangulation ranging with deep learning.

Area of Science:

• Optics and Photonics
• Artificial Intelligence
• Metrology

Background:

• Laser triangulation ranging systems are susceptible to measurement errors stemming from structural parameter inaccuracies.
• The asymmetric nature of these systems complicates precise calibration and parameter estimation.
• Existing methods often struggle with nonlinear errors and sensitivity to system design.

Purpose of the Study:

• To develop an intelligent laser triangulation ranging method that enhances measurement accuracy.
• To reduce the system's dependence on precise structural parameters.
• To improve the robustness and reliability of laser triangulation systems.

Main Methods:

• Utilized a nonlinear programming genetic algorithm with an elitist strategy for system design.
• Developed a deep neural network for distance estimation, bypassing traditional displacement equations.
• Implemented a deep learning approach to directly estimate distance from sensor data.

Main Results:

• Reduced nonlinear errors by an order of magnitude compared to conventional methods.
• Achieved sub-micron root mean squared error (RMSE) and mean absolute error (MAE).
• Demonstrated significantly lower RMSE (factor of 5) and MAE (factor of 3).

Conclusions:

• The proposed deep learning method substantially increases measurement accuracy in laser triangulation.
• This intelligent approach mitigates the impact of design and setting errors in system parameters.
• The technique offers a more robust and accurate solution for laser triangulation ranging applications.