Polarized Image Normal Line Calculation for Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies face challenges in accurately calculating normal lines from polarized images due to 180-degree periodicity issues and inability to identify three-dimensional shapes using only two polarization directions, and require high distance resolution images which are difficult to acquire.
Innovation Solution
An image processing apparatus and method that acquires polarized images with multiple polarization directions, calculates normal lines, and resolves uncertainties using temporary recognition processes and pre-registered models to determine object shapes and orientations accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If normal lines are calculated from polarized images with multiple polarization directions, then the brightness information is improved, but 180-degree uncertainty remains in the azimuth angle determination
Solution Approach 1:
The patent introduces a temporary recognition process as an intermediary step. This process generates a temporary recognition process image from the polarized images and uses it to temporarily recognize the object, which then serves as a mediator to resolve the 180-degree uncertainty in the normal line calculation by providing contextual information about the object's identity and orientation
Solution Approach 2:
The patent performs preliminary object recognition using a temporary recognition process before finalizing the normal line calculation. This preliminary action of temporarily recognizing the object based on the temporary recognition process image allows the system to prepare the necessary information (such as expected object orientation and identity) that will be used to resolve the azimuth angle uncertainty in the subsequent normal line calculation step
2Device complexity
If only two polarization directions are used, then the device complexity is reduced, but the ability to identify three-dimensional shape is insufficient
Solution Approach 1:
The patent transitions from using only intensity information to utilizing polarization direction information as an additional dimension. By calculating normal lines from the polarization characteristics of reflected light across multiple polarization directions, the system extracts three-dimensional shape information without requiring complex multi-camera or structured light systems, effectively adding a dimensional aspect to the measurement
3Measurement precision
If high distance resolution images are used for accurate target detection, then the measurement precision is improved, but the difficulty of acquiring such images increases
Solution Approach 1:
The patent replaces mechanical or complex imaging systems (such as high-resolution depth cameras or structured light systems) with a polarization-based optical measurement approach. By utilizing the polarization characteristics of light reflected from the target object and applying physical models of polarization reflection, the system achieves accurate distance and shape measurement without requiring mechanically complex high-resolution imaging equipment
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
Enables accurate recognition of objects with high precision by resolving normal line uncertainties and determining three-dimensional shapes from polarized images, improving upon existing limitations in object identification and shape determination.
Implementation Method 1
polarized images having a plurality of different polarization directions
Implementation Method 2
light reflected from the reference plane
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A polarized image acquisition section 20 acquires a plurality of polarized images having different polarization directions. The polarized images show, for example, an input indicator for a user interface as a recognition target object. A normal line calculation section 30 calculates normal lines for individual pixels of the recognition target object in accordance with the polarized images acquired by the polarized image acquisition section 20. The normal lines represent information based on the three-dimensional shape of the recognition target object. A recognition section 40 recognizes the object by using the normal lines calculated by the normal line calculation section 30, determines, for example, the type, position, and posture of the input indicator, and outputs the result of determination as input information on the user interface. The object can be recognized easily and with high accuracy.