Drone Camera Autoexposure Stabilization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Automatic exposure control algorithms in drones' cameras adjust based on the entire capture zone, including elements outside the final useful area, leading to overexposure issues due to changing sky-to-ground ratios during drone movements, affecting the final image quality.

Innovation Solution

The camera system analyzes the capture zone's image data to extract the raw useful area, applies image stabilization and rectification, and adjusts auto-exposure parameters by weighting regions of interest based on their overlap with the final useful area, ensuring accurate exposure settings for the user's viewed scene.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If the autoexposure control analyzes the entire capture zone image data, then it can control all elements in the captured scene, but it leads to overexposure because the analysis includes elements outside the final useful area which have different brightness characteristics

Engineering Contradiction:
Improvecontrol coverageVSAvoidexposure accuracy
Core Design Contradiction:
Adaptability or versatilityVSManufacturing precision

Solution Approach 1:

The patent segments the capture zone into multiple analysis zones with different weights. The useful area (corresponding to the final image after stabilization and rectification) is assigned a higher weight, while areas outside the useful area are assigned lower weights or excluded from analysis. This segmentation allows the autoexposure control to focus on the relevant scene elements that will actually appear in the final image, preventing overexposure caused by analyzing irrelevant areas.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by assigning different importance weights to different regions of the capture zone. Instead of uniform analysis across the entire capture zone, the system identifies the useful area that corresponds to the stabilized and rectified final image and gives it priority in the exposure calculation. This ensures that exposure settings are optimized for the actual visible content rather than including peripheral or unstable regions.

Inventive Principle:
Principle #3Local quality

2Stability of the object's composition

If the capture zone is stabilized by moving the analysis window in the opposite direction of drone movements, then image stability is improved, but the sky-to-ground ratio in the analysis area changes during attitude changes, causing exposure inconsistencies

Engineering Contradiction:
Improveimage stabilityVSAvoidexposure consistency
Core Design Contradiction:
Stability of the object's compositionVSManufacturing precision

Solution Approach 1:

The patent implements feedback by continuously monitoring the drone's attitude changes and dynamically adjusting the analysis zones and their weights accordingly. When the drone's attitude changes, the system recalculates which areas of the capture zone correspond to the useful area in the stabilized image and adjusts the exposure analysis weights to reflect the new sky-to-ground ratio. This feedback mechanism ensures exposure consistency despite changes in the stabilized analysis window position.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3142353B1Drone with forward-looking camera in which the control parameters, especially autoexposure, are made independent of the attitude
Publication Date: 2019.12.18 PARROT
  • EP3142353B1 patent drawingFigure 1~2b
  • EP3142353B1 patent drawingFigure 3(a)~3(b)
  • EP3142353B1 patent drawingFigure 4~5

AI summary

The drone comprises a camera (14), an inertial measurement unit (46) measuring the drone's angles, and an extraction module (52) delivering image data from a small, moving capture area that is dynamically displaced in the opposite direction to the angle changes measured by the inertial measurement unit. Compensating means (52) receive the drone's current attitude data as input and dynamically adjust the current value (54) of a shooting parameter such as autoexposure, white balance, or autofocus, calculated based on the image data contained in the capture area.