Pedestrian Detection Using Range Sensor Boundary Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional technologies face challenges in accurately detecting pedestrians in the surroundings of a vehicle, often leading to false detection of images or passengers as pedestrians, which can result in unnecessary vehicle deceleration.
Innovation Solution
A method that combines point group information from a range sensor and image recognition processing to identify pedestrians by extracting boundary position candidates based on distance changes and coinciding them with image-defined partial regions, preventing false detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If image recognition processing is used to detect pedestrians, then detection speed is improved, but false detection increases (images or passengers mistakenly detected as pedestrians)
Solution Approach 1:
The patent combines image recognition results with range sensor data to perform pedestrian detection. The image recognition unit provides candidate regions, while the range sensor verifies the physical presence and boundaries of objects. This merging of multiple detection methods maintains fast detection speed while improving accuracy by cross-validating results from different sensing modalities.
Solution Approach 2:
The patent introduces a boundary position candidate extraction unit that acts as an intermediary between image recognition and final pedestrian determination. This unit extracts boundary positions from range sensor data and compares them with image recognition results, serving as a verification layer that filters false positives while maintaining detection speed.
2Device complexity
If only image recognition is used for pedestrian detection, then device complexity is reduced, but detection precision deteriorates (false positives occur)
Solution Approach 1:
The patent divides the detection system into distinct functional units: image recognition unit for initial candidate identification, range sensor for physical verification, boundary position extraction unit for boundary analysis, and pedestrian determination unit for final decision. This segmentation allows each unit to perform its specialized function efficiently, improving overall precision without excessive complexity.
Solution Approach 2:
The patent adds a spatial dimension verification by using range sensor data to check the physical boundaries and distance of detected objects. While image recognition operates in 2D image space, the range sensor provides 3D spatial information, creating a new dimension of verification that significantly reduces false positives.
3Measurement precision
If boundary position comparison is performed to verify pedestrian detection, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary boundary position extraction from range sensor data before comparing with image recognition results. By preparing the boundary position candidates in advance and organizing them in a structured manner, the comparison process becomes more efficient, reducing the time penalty associated with enhanced accuracy verification.
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 enhances the precision of pedestrian detection, reducing false positives and ensuring accurate recognition of pedestrians in the vehicle's surroundings.
Implementation Method 1
a technology for extracting a group of pedestrian candidate points by grouping a group of points acquired by detecting a pedestrian by a laser radar
Implementation Method 2
the robot employs an imaging apparatus and a range detecting apparatus
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An object recognition method including: detecting a plurality of positions on surfaces of objects (100, 101) in surroundings of an own vehicle (1) along a predetermined direction and acquiring a group of points; generating a captured image of surroundings of the own vehicle; grouping points included in the acquired group of points and classifying the points into a group (p1 to p21) of object candidate points; extracting, from among object candidate points, the object candidate points being points included in the group of object candidate points, a position (p1, p7, p10, p21) at which change in distance from the own vehicle between adjacent object candidate points increases from a value equal to or less than a predetermined threshold value to a value greater than the predetermined threshold value as a boundary position candidate, the boundary position candidate being an outer end position of an object (100, 101); extracting a region (R) in which a person (100) is detected in the captured image as a partial region by image recognition processing; and when, in the captured image, a position of the boundary position candidate (p1, p7, p10, p21) coincides with a boundary position of the partial region (R), the boundary position being an outer end position, in the predetermined direction, recognizing that a pedestrian exists in the partial region (R) .