Optical Reflection Feature Matching for Low-Compute 3D Position Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical 3D sensing methods face challenges with unreliable results in environments with multiple reflections, high computational demands, and resource-intensive computational power requirements, limiting their use in mobile and outdoor applications.
Innovation Solution
A detector system utilizing a matrix of optical sensors with an evaluation device that determines object position by optimizing a blurring function to estimate distance and match reflection features, reducing computational complexity and resource demands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical 3D sensing methods are used to determine object position, then measurement precision is improved, but computational power requirements increase
Solution Approach 1:
The evaluation device segments the correspondence problem by first identifying candidate feature points in the first image, then for each candidate point, determining only the specific corresponding point on the epipolar line in the second image. This divides the complex global correspondence problem into smaller local matching tasks, reducing computational power requirements while maintaining measurement precision.
Solution Approach 2:
The method performs preliminary actions by pre-identifying candidate feature points and establishing epipolar lines before solving the correspondence problem. By preparing these structural elements in advance, the actual correspondence matching requires less computational power, yet achieves accurate object position determination.
2Device complexity
If structured light methods with fixed relative position detectors are used, then device complexity is reduced, but measurement precision deteriorates due to correspondence problem difficulties
Solution Approach 1:
The method introduces epipolar lines as an intermediary geometric constraint to solve the correspondence problem. By restricting the search for corresponding points to specific epipolar lines rather than the entire image, the method maintains measurement precision while working with the fixed relative position detector arrangement, thus not increasing device complexity.
3Measurement precision
If quadratic scaling evaluation algorithms are employed to solve correspondence problems, then measurement precision is improved, but use of energy increases
Solution Approach 1:
The evaluation algorithm is segmented to process only candidate feature points and their corresponding points on epipolar lines, rather than comparing all points in the image. This segmentation reduces the computational complexity from quadratic scaling to linear or near-linear scaling, significantly reducing energy consumption while preserving measurement precision through careful candidate selection and epipolar line constraints.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system provides reliable object position determination with low technical effort and resource requirements, enhancing usability in mobile and outdoor environments.
Implementation Method 1
at least one sensor element having a matrix of optical sensors, the optical sensors each having a light-sensitive area, wherein the sensor element is configured to determine at least one reflection image of the object
Data Source
AI summary
Described herein is a detector for determining a position of at least one object. The detector includes at least one sensor element having a matrix of optical sensors, the optical sensors each having a light-sensitive area, wherein the sensor element is configured to determine a reflection image of the object. The detector also includes an evaluation device configured to select a reflection feature of the reflection image, and determine a distance estimate of the selected reflection feature of the reflection image by optimizing at least one blurring function fa, wherein the distance estimate is given by a longitudinal coordinate z and an error interval±ε. The evaluation device is adapted to determine at least one displacement region in at least one reference image corresponding to the distance estimate, and to match the selected reflection feature with at least one reference feature within the displacement region.