Monocular Depth Map Scaling with Pixel Ratio Distributions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing depth maps created from camera images using monocular optical systems lack a consistent scale, leading to discontinuities when fused, as they rely on relative depth information and are sensitive to camera assembly position changes, necessitating a method to align and scale these maps accurately.
Innovation Solution
A method involving pixel-level ratio calculation between depth values, frequency distribution analysis, and determination of a global scaling factor using statistical parameters, particularly Gaussian mixing models, to align and fuse depth maps effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If depth maps are created using monocular optical systems with relative depth information, then depth information can be obtained from camera images, but the scale of the depth maps is inconsistent leading to discontinuities in fused results
Solution Approach 1:
The patent applies parameter changes by calculating a scaling factor for each depth map based on the statistical properties (mean and standard deviation) of depth values in reference regions. This transforms the depth maps from having inconsistent scales to having a unified scale, resolving the contradiction between obtaining depth information and maintaining scale consistency.
2Adaptability or versatility
If different algorithms are used to create depth maps from the same camera image, then multiple depth maps can be generated, but the scaling between depth maps varies causing fusion discontinuities
Solution Approach 1:
The patent implements feedback by using the statistical properties of depth values from reference regions to calculate and adjust the scaling factors of depth maps. This feedback mechanism ensures that regardless of which algorithm is used to generate a depth map, the resulting scale can be corrected to match the reference scale, maintaining both versatility and precision.
3Adaptability or versatility
If camera assembly position changes occur, then depth information can be updated, but the scaling of depth maps becomes sensitive to position changes requiring realignment
Solution Approach 1:
The patent applies preliminary action by establishing a reference region with known scaling properties before processing depth maps from different camera positions. By pre-defining the reference scale, the system can reliably calculate scaling factors for new depth maps regardless of camera position changes, ensuring both adaptability and scaling stability.
Data Source
AI summary
Method for ascertaining at least one scaling factor for a first depth map (1), said method having the following steps:a) extracting first depth values (3) d1,i for a plurality of pixels i of the first depth map;b) ascertaining a comparison depth value (4) d2,i for each first depth value (3) d1,i;c) calculating ratiosSr,i=d1,id2,i between the first depth value (3) d1,i and the comparison depth value (4) d2,i for each pixel i;d) ascertaining a frequency distribution (5) of the ratio Sr,i over all pixels i; ande) determining the at least one scaling factor on the basis of the frequency distribution (5).