Underwater optical signal detection system using diffuser-based lensless imaging.

Area of Science:

• Optics and Photonics
• Signal Processing
• Machine Learning

Background:

• Underwater optical communication faces challenges from turbid media and occlusion.
• Traditional lens-based systems can be bulky and computationally expensive.
• Developing robust and cost-effective underwater signal detection is crucial.

Purpose of the Study:

• To propose and evaluate a novel diffuser-based lensless underwater optical signal detection system.
• To compare the performance and computational cost against a lens-based system.
• To investigate the impact of dimensionality reduction on lensless system performance.

Main Methods:

• Utilizing a lensless one-dimensional (1D) camera array with random phase modulators for signal acquisition.
• Employing a one-dimensional integral imaging convolutional neural network (1DInImCNN) for signal classification.
• Transmitting signals through a turbid medium and partial occlusion during data capture.

Main Results:

• The proposed lensless system demonstrated superior detection performance compared to an equivalent lens-based system.
• Computational costs were analyzed and found to be favorable for the lensless approach.
• Dimensionality reduction on lensless patterns showed minimal impact on detection performance.

Conclusions:

• Diffuser-based lensless systems offer a viable and high-performing alternative for underwater optical signal detection.
• These systems are well-suited for low-cost compressive underwater optical imaging applications.
• The lensless approach provides a robust solution for signal recovery in challenging underwater environments.