POLARX PHOTONICS

Case Study

Filter-array bonding, calibration, and ISP integration

IMX990 SWIR Polarization Camera

PolarX Photonics aligns and bonds pixelated filter arrays to arbitrary image sensors, then develops the calibration and ISP needed for usable system output. This case study summarizes a Sony IMX990-based SWIR polarization camera built through that workflow.

Sony IMX990 SWIR imaging Pixelated filter arrays Precision bonding

Case-study platform

Sony IMX990-AABJ

SVS-VISTEK exo990MGE

PolarX scope

Precision alignment and bonding, internal process software, calibration, customer software / ISP, and deployment support

PolarX process environment with realtime alignment display monitors
Realtime sensor feedback during alignment optimization.

Workflow

Platform-specific pixelated array integration workflow

Each camera platform requires its own mechanical, optical, and software path. The workflow below summarizes how PolarX adapts pixelated filter arrays to a given sensor platform, including systems that require cover-glass removal before bonding.

01

Platform review and sensor intake

Sensor stack, package geometry, optical path, and camera constraints are reviewed to define the integration plan and risk points for the target platform.

02

Cover-glass removal process development

If the cover glass is not removable, PolarX develops dedicated removal tooling and SOPs using a UV/CO2 laser ablation system, followed by cleaning and preparation.

03

Camera-platform-specific tooling

Custom fixtures and camera-platform-specific tooling are developed for handling, registration, bonding, and any required rework on the target hardware.

04

Filter-array integration process

The bonding process is defined for the selected pixelated filter array, including adhesive stack, placement strategy, and process controls for repeatable integration.

05

Cleanroom alignment and bonding

Alignment is performed in a cleanroom using realtime sensor feedback, ROI metrics, and process control to push registration as far as the sensor geometry allows.

06

Post-alignment performance verification

After bonding, optical and mechanical performance verification is performed to confirm registration, inspect the assembly, and benchmark the bonded hardware baseline.

07

Calibration and ISP development

Calibration files and ISP processing are developed and tuned to convert the bonded stack into stable outputs and improve delivered effective contrast.

08

Custom software integration

Customer software integration, visualization, controls, and export workflows are developed for the target camera platform and application.

Software Stack

Process and application software

The first interface shown is internal-use process software for alignment and bonding, using realtime sensor feedback to optimize placement before final attachment. The second is customer-facing camera software and ISP tooling, which PolarX can develop and customize for the target system.

Polar alignment metrics software showing QC regions and maps

Internal use only

Realtime alignment feedback environment

This environment is used during process development and assembly. It uses realtime sensor feedback, ROI metrics, and spatial checks to drive alignment toward the best physically attainable registration.

Polarization ISP software showing calibration and display tuning

Customer software / customizable

Calibration-aware camera software and ISP

Calibration loading, visualization, tuning controls, and export tools can be developed and customized for the customer workflow and target application.

Live Demo

Sensor output in action

The same platform can deliver synchronized polarization-derived views such as AoP, DoLP, intensity, and HSV fusion within the software stack.

IMX990 live demo output showing AoP, intensity, DoLP, and HSV fusion views

Performance Review

Raw and calibrated data should be read together

The raw plots show the bonded hardware baseline after physical alignment has been optimized. The calibrated plots show the additional system-level gain from calibration files and ISP processing.

Raw extinction ratio spectral response plot

Raw extinction ratio

Baseline ER of the bonded sensor and filter stack before calibration files and ISP are applied.

Calibrated effective extinction ratio spectral response plot

Calibrated effective ER

Calibration files and ISP raise delivered effective ER and contrast ratio on the bonded hardware baseline.

Raw Malus response plot at 600 nm

Raw Malus response @ 600 nm

The raw response shows the bonded hardware baseline with usable four-channel sinusoidal separation.

Calibrated channel response plot at 600 nm

Calibrated channel response @ 600 nm

Calibration files and ISP preserve channel separation while improving the delivered system response.

Platform Extension

Applicable beyond polarization cameras

This IMX990 case study is one example of a broader process. The same bonding, calibration, and software approach can be adapted to other sensors, pixelated filter arrays, and customer-specific imaging pipelines.

RGBN pixelated filter array example

RGBN pixelated filter array example

Example of a visible-band pixelated filter array architecture beyond polarization imaging.

Visible and NIR spectral response example

Visible + NIR spectral response example

Example broadband response spanning visible and near-infrared operation for non-polarization filter-array systems.

Polarization arrays Multispectral mosaics Custom pixelated filter arrays

Contact

Technical contact

For technical follow-up, program discussion, or application review, please contact:

Jonathan Zhu

Co-Founder | Imaging Systems & Photonics Engineering

PolarX Photonics

Champaign, IL, USA