Monocular Depth Reliability Mapping Using Ground Region Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Monocular depth estimation in images, particularly in automated driving scenarios, suffers from accuracy deterioration when the ground portion between a subject and the ground is unclear due to occlusion, leading to potential collisions with obstacles.

Innovation Solution

An image processing apparatus that includes a depth information acquisition unit, a ground region determination unit, and a reliability calculation unit to determine the ground region and assign reliability to the depth information, enhancing accuracy by segmenting the image and assigning high reliability to the ground region.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If monocular depth estimation is used to acquire depth information from images, then the system can estimate depth using simple hardware, but the accuracy deteriorates when the ground portion between subject and ground is unclear due to occlusion

Engineering Contradiction:
Improvehardware simplicityVSAvoiddepth estimation accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The image is segmented into multiple regions including ground regions and non-ground regions. By dividing the image into distinct segments, the system can apply different reliability assessments to different areas, identifying regions where occlusion may have affected depth estimation accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different reliability values are assigned to different regions of the image based on local characteristics. Ground regions are identified and given appropriate reliability weights, allowing the system to account for local variations in measurement quality without compromising the overall depth estimation.

Inventive Principle:
Principle #3Local quality

2Reliability

If reliability calculation is performed based on ground region determination, then depth estimation reliability improves, but the processing complexity increases

Engineering Contradiction:
Improvedepth estimation reliabilityVSAvoidprocessing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

Ground region determination is performed as a preliminary step before final depth estimation. By pre-identifying ground regions and assigning reliability values in advance, the system prepares reliability information that can be efficiently applied during subsequent processing stages.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

A reliability map is introduced as an intermediary data structure that stores reliability information for different image regions. This intermediary layer allows the system to decouple the complexity of ground region analysis from the depth estimation process, enabling efficient integration of reliability information.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS20260075177A1Image processing apparatus, image processing method, and storage medium
Publication Date: 2026.03.12 CANON KK
  • US20260075177A1 patent drawing
  • US20260075177A1 patent drawing
  • US20260075177A1 patent drawing

AI summary

An image processing apparatus that obtains a depth from an input image includes: a depth information acquisition unit configured to acquire depth information; a ground region determination unit configured to determine a ground region of a subject of the image; and a reliability calculation unit configured to calculate reliability for the depth acquired by the depth information acquisition unit based on the determination of the ground region determination unit. Here, the depth is an absolute distance between a standard point and a subject or an evaluation value relatively indicating a distance between a standard point and a subject. The reliability calculation unit calculates high reliability for a depth of a pixel of the ground region or a segmented region including the ground region.