Camera Speed Estimation via Pixel Intensity Minimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating the speed of a video camera while capturing a 3D scene are complex and prone to noise, especially when dealing with rolling shutter distortion and motion blur, and often require feature extraction and matching steps.
Innovation Solution
A method that directly minimizes differences in physical magnitudes across a large number of pixels in both reference and current images, eliminating the need for feature extraction and matching, and accounts for camera movement to improve estimation accuracy, even in the presence of rolling shutter and motion blur distortions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature-based methods are used to estimate camera speed, then the estimation can be performed using traditional image processing techniques, but the method becomes complex and prone to noise due to feature extraction and matching steps
Solution Approach 1:
The patent extracts and eliminates the complex feature extraction and matching steps from the traditional pipeline. Instead of using features as intermediaries, the method directly uses pixel intensity values from the image grid to estimate camera speed, thereby removing the source of complexity and noise while maintaining estimation capability
Solution Approach 2:
The patent replaces the mechanical/image-processing system of feature detection and matching with a direct mathematical optimization approach. By formulating speed estimation as a minimization problem using pixel intensity differences, the method substitutes complex computational image processing with a more robust algebraic solution
2Reliability
If feature extraction and matching steps are used, then traditional image processing techniques can be applied, but the method is affected by noise and not robust
Solution Approach 1:
The patent converts the typically harmful effect of motion blur into a useful signal. By modeling the motion blur as a convolution operation and incorporating it into the optimization framework, the method transforms what was previously a distortion to be corrected into information that helps estimate camera speed and motion parameters
Solution Approach 2:
The patent changes the fundamental parameters used for estimation from discrete feature points to continuous pixel intensity values across the entire image grid. This parameter transformation increases the amount of data available for estimation (from a handful of features to thousands of pixels) and makes the method more robust to noise through statistical averaging
3Productivity
If rolling shutter video cameras are used, then the capture speed can be increased, but distortion appears in the captured image due to temporal offset between pixel rows
Solution Approach 1:
The patent performs preliminary modeling of the rolling shutter distortion effect before attempting to correct it or use it for estimation. By incorporating the known temporal offset pattern into the optimization model, the method prepares for and compensates for the distortion inherently, rather than trying to remove it afterward
Solution Approach 2:
The patent converts the rolling shutter distortion, typically considered a harmful artifact, into a useful source of information. The systematic temporal offset pattern encodes camera motion information that can be extracted to estimate speed and correct the distortion, turning a limitation into an advantage
4Measurement precision
If exposure time is increased to improve image quality, then luminous intensity measurement improves, but motion blur increases when the camera moves during exposure
Solution Approach 1:
The patent changes the approach from treating motion blur as a degradation to be minimized by reducing exposure time, to treating it as a controllable parameter that provides motion information. By incorporating the exposure time and camera speed into the optimization model, the method can use longer exposures beneficially while accurately estimating and compensating for the resulting blur
Data Source
AI summary
A method for estimating the speed of movement of a first video camera when it captures a current image of a three-dimensional scene, includes storing a reference image corresponding to an image of the same scene captured by a second video camera in a different pose, the reference image including pixels. The method also includes storing the current image, the current image including pixels containing the measurement of a physical quantity measured by that pixel, that physical quantity being the same as the physical quantity measured by the pixels of the reference image. The method further includes storing for each pixel of the reference image or of the current image the measurement of a depth that separates that pixel from the point of the scene photographed by that pixel, estimating the pose of the first camera, and estimating the speed of movement of the first camera.