Laser Pattern Image Feature Extraction for Short-Range Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for distance measurement and object detection in autonomous driving, such as using three-dimensional depth cameras, are costly and time-consuming, especially when accurate depth information is not required for objects close to the robot.
Innovation Solution
A method using a laser pattern to extract features from images, where a first camera captures a laser pattern reflected from objects and a second camera captures overlapping areas, with a controller generating a mask to distinguish effective areas and extract features, thereby reducing processing time and identifying objects at short distances.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a three-dimensional depth camera is used to measure distance and detect objects, then measurement precision and object detection capability are improved, but device complexity and cost increase
Solution Approach 1:
The patent introduces a laser pattern as an intermediary element projected onto objects in the scene. Instead of using complex depth cameras, the system projects structured light patterns and uses a standard 2D camera to capture the reflected patterns. The laser filter acts as a mediator to isolate the laser pattern from other light sources, enabling depth information extraction through pattern analysis rather than direct depth sensing.
Solution Approach 2:
The patent replaces the mechanical/optical depth camera system with a laser projection and 2D camera capture system. By substituting the depth measurement mechanism with laser pattern projection and analysis, the system achieves depth information extraction using simpler, more cost-effective components while maintaining measurement capability.
2Measurement precision
If a three-dimensional depth camera is used for object detection in autonomous driving, then object detection accuracy is improved, but arithmetic operation time increases
Solution Approach 1:
The system performs preliminary action by projecting the laser pattern onto objects before capturing the image. This pre-structuring of light patterns on objects enables the 2D camera to capture depth-encoded information directly in the captured image, eliminating the need for complex post-capture depth calculation algorithms and reducing arithmetic operation time.
Solution Approach 2:
The patent utilizes the laser pattern's distinct optical characteristics (analogous to color changes) by employing a laser filter to isolate the laser wavelength from other light sources. This spectral separation allows the system to identify and process only the laser-reflected patterns, simplifying image processing and reducing computational overhead for object detection.
3Reliability
If feature extraction is performed on all objects in the image, then comprehensive object identification is achieved, but processing time increases
Solution Approach 1:
The patent extracts only the relevant laser pattern information from the captured image by using a laser filter to isolate the laser wavelength. This extraction process removes irrelevant information (other light sources, background elements) and focuses processing only on the laser-reflected patterns, thereby reducing feature matching time while maintaining identification reliability for objects within the effective range.
Solution Approach 2:
The system applies local quality by generating a mask that distinguishes the effective area (where laser patterns are present) from the ineffective area. Feature extraction and matching are performed only within the effective area defined by the mask, rather than processing the entire image. This localized processing significantly reduces computation time while maintaining comprehensive identification of objects in the relevant range.
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
This approach reduces feature matching time and allows for efficient identification of objects at short distances by using a laser pattern to remove objects at long distances from the image, improving the speed and accuracy of the feature extraction process.
Implementation Method 1
a first camera coupled to a laser filter and configured to generate a first image including a pattern of a laser which is reflected from an object
Implementation Method 2
a pattern of a laser which is reflected from an object
Data Source
AI summary
Provided herein are a method of extracting a feature from an image using a laser pattern and an identification device and a robot including the same, and the identification device for extracting a feature from an image using a laser pattern, which includes a first camera coupled to a laser filter and configured to generate a first image including a pattern of a laser which is reflected from an object, a second camera configured to capture an area overlapping an area captured by the first camera to generate a second image, and a controller configured to generate a mask for distinguishing an effective area using the pattern included in the first image and extract a feature from the second image by applying the mask to the second image.