Application of Zernike polynomial optimized lens design in orientation detection

This paper employs the Zernike polynomials to optimize lens design, effectively overcoming the limitations of the traditional Simultaneous Multiple Surface (SMS) method in achieving uniform and symmetric light spot, thereby improving optical performance. This study constructs a non-imaging optical system by integrating the optimized lens with a detector array and applying the random forest algorithm to achieve orientation detection and prediction. The results show that the system’s azimuth and pitch field of view(FOV) are both ±6°, with a resolution 1°. The algorithm’s mean squared error (MSE) is 0.1593, and the determination coefficient (R²) is 0.9854, indicating good robustness. The system demonstrates broad application potential and market prospects in high-precision orientation detection.