Polarization Imaging and Edge Detection with Image-Processing Metasurfaces

Michele Cotrufo,Sahitya Singh,Akshaj Arora,Alexander Majewski,Andrea Alù

DOI: https://doi.org/10.1364/OPTICA.500121

2023-07-07

Abstract:Optical metasurfaces have been recently explored as ultrathin analog image differentiators. By tailoring the momentum transfer function, they can perform efficient Fourier-filtering - and thus potentially any linear mathematical operation - on an input image, replacing bulky 4f systems. While this approach has been investigated in different platforms, and several techniques have been explored to achieve the required angular response, little effort has been devoted so far to tailor and control also the polarization response of an image-processing metasurface. Here, we show that edge-detection metasurfaces can be designed with tailored polarization responses while simultaneously preserving an isotropic response. In particular, we demonstrate single-layer silicon metasurfaces yielding efficient Laplacian operation on a 2D image with either large polarization asymmetry, or nearly polarization-independent response. In the former case, we show that a strongly asymmetric polarization response can be used to unlock more sophisticated on-the-fly image processing functionalities, such as dynamically tunable direction-dependent edge detection. In parallel, metasurfaces with dual-polarized response are shown to enable efficient operation for unpolarized or arbitrarily polarized images, ensuring high efficiency. For both devices, we demonstrate edge detection within relatively large numerical apertures, with excellent isotropy and intensity throughput. Our study paves the way for the broad use of optical metasurfaces for sophisticated, massively parallel analog image processing with zero energy requirements.

Optics,Applied Physics

What problem does this paper attempt to address?

The problem this paper attempts to address is achieving efficient edge detection and image processing by designing metasurfaces with specific polarization responses. Specifically, the authors aim to solve the following two main issues:

1. **Control of Polarization Response**:
- **Existing Problem**: Most current metasurfaces used for edge detection primarily focus on angular response during design, with less control over polarization response. This limits the functionality and efficiency of these metasurfaces in practical applications.
- **Objective**: Design metasurfaces that can control both polarization response and angular response simultaneously to achieve more complex and flexible image processing functions.

2. **Efficient and Isotropic Edge Detection**:
- **Existing Problem**: Existing edge detection metasurfaces have deficiencies in efficiency and isotropy (i.e., uniform response in all directions) when processing input light with different polarization states.
- **Objective**: Develop metasurfaces capable of achieving polarization-dependent or polarization-independent edge detection while maintaining high efficiency and isotropy.

### Specific Problems and Solutions

1. **Polarization-Dependent Edge Detection**:
- **Problem**: How to design a metasurface that enhances edges in a specific direction while ignoring edges in other directions?
- **Solution**: By designing metasurfaces with strong polarization asymmetric responses, only edges parallel to the input polarization direction are enhanced. For example, when the input light is horizontally polarized, only horizontal edges are enhanced.

2. **Polarization-Independent Edge Detection**:
- **Problem**: How to design a metasurface that can efficiently perform edge detection for input light with any polarization state?
- **Solution**: By optimizing the design parameters of the metasurface, making the transmission functions for s-polarization and p-polarization similar, thus achieving efficient and isotropic edge detection for input light with any polarization state.

### Experimental Verification

- **Experimental Results**: The authors experimentally verified the designs of these two types of metasurfaces. For the polarization-dependent metasurface, the experimental results showed that only edges parallel to the input polarization direction were enhanced, while edges in other directions were not. For the polarization-independent metasurface, the experimental results showed that edges in all directions were uniformly enhanced regardless of the polarization state of the input light.

### Conclusion

Through the above research, the authors demonstrated how to achieve efficient and flexible edge detection and image processing by controlling the polarization response of metasurfaces. This achievement provides new ideas and methods for future optical image processing technologies.