Microlens Array

The pitch is the center-to-center distance between adjacent microlenses, defining the spatial sampling rate of the light field. A smaller pitch increases spatial resolution but reduces the number of angular samples per lens. The fill factor is the percentage of the sensor area covered by active lens elements; a high fill factor (often >90%) maximizes light collection efficiency and minimizes optical crosstalk between adjacent sampling units.

The focal length of each individual microlens determines where the main lens's pupil is imaged onto the photosensor. This parameter sets the angular resolution—the number of distinct ray directions sampled per spatial point. The f-number (focal length / aperture diameter) of the microlenses affects the system's light-gathering ability and depth of field within each micro-image. Typical microlens focal lengths range from tens to hundreds of micrometers.

Microlenses are fabricated with specific surface profiles to control optical performance. Common profiles include:

• Spherical: Simplest to manufacture but introduces spherical aberration.
• Aspherical: Corrects for spherical aberration, improving image quality at the micro-image plane.
• Diffractive: Uses nanostructures to achieve specific focusing properties, often for chromatic correction. Minimizing chromatic aberration, coma, and astigmatism is critical to ensure each micro-image is a sharp, accurate sample of the incident light cone.

This is the fundamental constraint governing microlens array design. For a fixed sensor with N total pixels:

• High Spatial Resolution: Achieved by using a large number of microlenses (small pitch), each covering few sensor pixels. This yields a high-resolution 2D image but very coarse angular information.
• High Angular Resolution: Achieved by using fewer, larger microlenses (large pitch), each covering many sensor pixels. This provides rich directional light data but at the cost of low spatial resolution in the final rendered image. System design involves optimizing this tradeoff for the target application, such as refocusing vs. 3D reconstruction.

Microlens arrays are typically fabricated from optical-grade polymers (e.g., PMMA) or fused silica. Key material properties include:

• Refractive Index: Determines the lens's bending power for a given curvature.
• Dispersion: Affects chromatic performance; low dispersion is preferred.
• Thermal Stability: Critical for systems operating in varying environmental conditions. The array is often bonded directly to the sensor's protective glass (cover slip) or fabricated as a separate component precisely aligned to the pixel grid. The substrate thickness and flatness are tightly controlled to maintain optical path consistency.

Precise alignment between the microlens array and the underlying photosensor pixel grid is paramount. Misalignment causes systematic errors in ray assignment, degrading light field reconstruction. Telecentricity refers to the property where the chief rays (central rays of each light cone) are parallel to the optical axis. A telecentric microlens array design ensures each microlens samples the same angular range from the main lens, simplifying calibration and post-processing algorithms for applications like the Lytro or Raytrix plenoptic cameras.

The quintessential application is in light field cameras (e.g., Lytro, Raytrix). The array is placed at the sensor's focal plane, where each microlens directs light from different angles onto distinct sensor pixels. This captures the 4D light field, enabling post-capture capabilities like:

• Digital refocusing: Changing the focal plane after the photo is taken.
• Parallax-based depth estimation: Computing depth maps from the angular disparity of rays.
• Perspective shift: Generating slightly different viewpoints from a single capture.

In astronomy and microscopy, microlens arrays are used in Shack-Hartmann wavefront sensors. Each lenslet creates a focal spot on a detector; deviations of these spots from a reference grid directly measure local wavefront tilt. This data drives deformable mirrors in adaptive optics systems to correct for atmospheric turbulence or optical aberrations in real time, dramatically improving image resolution.

Arrays are used to transform non-uniform light sources into uniform, well-defined illumination patterns. Each microlens segments the beam, and the overlapping outputs create a top-hat intensity profile. This is critical in:

• Photolithography: Providing even illumination on semiconductor wafers.
• Projection systems: Enhancing brightness uniformity.
• Microscopy: Creating uniform Köhler illumination for brightfield imaging.

Microlens arrays enable autostereoscopic 3D displays without requiring glasses. In integral imaging, a display panel shows a series of elemental images (different perspective views). The overlying microlens array directs light from each pixel to a specific direction, reconstructing a full-parallax light field viewable from multiple angles. This principle is also used in holographic stereogram printing.

Arrays facilitate high-precision measurement. In wavefront coding, a specialized microlens array creates a depth-invariant point spread function, extending depth of field for machine vision inspection. They are also used in moire interferometry and structured light projection for surface contour mapping, where the array generates precise grid or dot patterns for 3D scanning.

Inspired by insect compound eyes, engineered microlens arrays create multi-aperture imaging systems. Each lens feeds a small, independent sensor region, enabling novel form factors and capabilities:

• Ultra-thin cameras: For mobile devices.
• Wide-field-of-view imaging: By arranging lenses on a curved surface.
• Specialized sensing: Different lenslets can have varied focal lengths or filters for multispectral imaging or polarization analysis in a single compact module.

A light field is a 5D vector function that describes the radiance of light rays traveling in every direction (2D) through every point (3D) in space. It is a practical, finite subset of the full 7D plenoptic function, often parameterized for capture and rendering. In a plenoptic camera, the microlens array samples this 5D function, trading spatial resolution for angular information to enable post-capture effects like refocusing and parallax generation.

A plenoptic camera (or light field camera) is an imaging device that captures both the intensity and direction of light rays. Its core optical design places a microlens array between the main lens and the photosensor. Each microlens directs a bundle of rays from different angles onto a small patch of pixels beneath it. This raw capture, called a light field image, encodes the 4D light field, enabling computational photography applications impossible with conventional cameras.

Sub-aperture images are a set of 2D perspective views extracted during the decoding of a light field captured by a microlens array. Each image corresponds to the scene as seen from a different, small portion of the camera's main aperture.

• They represent different viewpoints with slight parallax.
• The complete set provides the angular sampling of the light field.
• These images are the primary data used for stereo matching, depth estimation, and view synthesis in light field processing pipelines.

The spatial-angular tradeoff is a fundamental resolution constraint in microlens-based light field capture. For a fixed sensor pixel count, the total information is divided between:

• Spatial Resolution: The number of distinct microlenses, determining the pixel count of each extracted sub-aperture image.
• Angular Resolution: The number of pixels behind each microlens, determining the number of unique sub-aperture images (views). Increasing one necessitates decreasing the other. This tradeoff drives hardware design and motivates super-resolution algorithms in computational light field photography.

Integral imaging is a passive, lenslet-based technique for capturing, processing, and displaying full-parallax 3D visual information. It directly employs a microlens array (or pinhole array) both at capture and display.

• Capture: Identical to a plenoptic camera, the array samples the light field.
• Display: A second microlens array reconstructs the light field, directing different image elements to each eye to create an autostereoscopic 3D effect without special glasses. It is a core technology for advanced 3D displays.

An Epipolar Plane Image is a powerful 2D visualization and analysis tool derived from a 4D light field. It is created by fixing one spatial dimension and one angular dimension, resulting in a 2D slice (y, v) or (x, u). Within an EPI:

• Points at different depths manifest as lines with distinct slopes.
• The slope of a line is directly proportional to the depth of the corresponding scene point.
• This linear structure enables highly efficient depth estimation and light field super-resolution algorithms by analyzing line orientations in this transformed domain.