Structured Differential Motion Estimation for HDR LED Flicker Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing systems struggle to differentiate between LED flicker and moving objects in HDR imagery, leading to inaccurate image capture and recognition, particularly in environments with both LED light sources and moving objects, due to the complexity of pinpointing LED flicker location and limitations of current LED flicker mitigation solutions.

Innovation Solution

A structured differential motion estimation algorithm using spatially multiplexed image sensors and a four-step workflow to distinguish between LED flickers and moving objects by analyzing pixel intensity patterns, including differential calculation, motion position estimation, structured motion search, and motion weighting estimation, to generate optimal HDR images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If LED flicker mitigation is applied to all pixels, then LED flicker artifacts are reduced, but moving objects become blurred due to incorrect processing

Engineering Contradiction:
ImproveLED flicker mitigation accuracyVSAvoidmoving object detection accuracy
Core Design Contradiction:
ReliabilityVSManufacturing precision

Solution Approach 1:

The patent segments the image into different regions: LED flicker regions and moving object regions. By calculating motion vectors and comparing them against LED flicker characteristics, the system identifies which pixels correspond to LED sources and which correspond to moving objects. This segmentation allows different processing strategies to be applied to different regions, resolving the contradiction between mitigating LED flicker and preserving moving object clarity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by treating different spatial regions of the image differently. Instead of uniformly applying LED flicker mitigation to all pixels, the system selectively applies processing based on local motion characteristics. Pixels identified as part of moving objects receive different treatment than pixels identified as LED sources, thereby preserving moving object detail while still mitigating LED flicker in appropriate regions.

Inventive Principle:
Principle #3Local quality

2Reliability

If camera exposure is synchronized with LED PWM frequency, then LED flicker artifacts are eliminated, but the system becomes complex and not universally applicable

Engineering Contradiction:
ImproveLED flicker artifact reductionVSAvoidsynchronization system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent replaces the mechanical/synchronization-based approach with a computational/image-processing-based approach. Instead of requiring hardware synchronization between camera exposure and LED PWM frequency, the system uses post-capture image processing to identify and mitigate LED flicker artifacts. This substitution eliminates the need for complex synchronization mechanisms while achieving similar or better results.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the approach from controlling exposure timing parameters to analyzing and processing image data parameters. By working with captured image data rather than controlling camera exposure timing, the system avoids the complexity of synchronization while still能够有效 addressing LED flicker. The motion vector calculation and comparison parameters enable flexible, parameter-driven LED flicker mitigation without hardware synchronization.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If motion estimation is performed on all pixels, then moving objects are accurately tracked, but processing time increases significantly

Engineering Contradiction:
Improvemoving object tracking accuracyVSAvoidimage processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies segmentation by first performing motion estimation on a subset of pixels or using efficient algorithms to identify regions of interest. By focusing computational resources on areas with significant motion or potential LED flicker, rather than uniformly processing all pixels, the system maintains tracking accuracy while reducing overall processing time.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements partial action by performing full motion estimation only where necessary (e.g., in regions with detected changes or potential LED sources) rather than on the entire image. This selective approach applies sufficient processing to maintain accuracy in critical regions while avoiding excessive processing in static or less important areas, thereby optimizing the balance between tracking accuracy and processing time.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS20250272792A1Light-emitting diode (LED) flicker detection based on structured differential motion estimation under high dynamic range (HDR) images
Publication Date: 2025.08.28 CISTA SYST
  • US20250272792A1 patent drawing
  • US20250272792A1 patent drawing
  • US20250272792A1 patent drawing

AI summary

This application describes an image processing system and method for LED Flicker Mitigation (LFM). An example method involves capturing multiple frames of a scene using an image sensor configured with various exposure settings. Based on the pixel values in the frames, the edges of a changing object may be identified in each frame, which could be a moving object or an LED source. It then determines the object's moving path between two frames captured with different exposure settings. The amount of movement of the changing object is calculated based on this moving path. The system then estimates the likelihood of the changing object being either the moving object or the LED source, based on the amount of movement. Finally, an image of the scene is generated using the captured frames by performing High Dynamic Range (HDR) fusion and LFM, taking into account the likelihood of the changing object's identity.