Vehicle Door Controller for Partial Intrusion Depth Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing door control systems struggle to accurately determine if a person has intruded into a camera's view, especially when only part of the person is visible, leading to potential safety issues.
Innovation Solution
A door controller using machine learning to analyze imaging-data by recognizing user regions and calculating depth differences between user and intrusion determination regions, enhancing accuracy in determining intrusions through both overlapping and depth comparisons.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image comparison methods are used to detect intruding objects, then the system can identify complete objects in the camera view, but it fails to detect partial intrusions where only part of a person enters the imaging area
Solution Approach 1:
The patent segments the detection task into two independent analysis dimensions: (1) user region recognition that identifies any part of a person in the image, and (2) depth difference calculation that measures distance from the door. This segmentation allows the system to detect partial intrusions by analyzing local regions rather than requiring complete object detection, thereby resolving the contradiction between detection accuracy and partial object capability.
Solution Approach 2:
The patent introduces a depth dimension by calculating depth differences between the user region and the door plane. This transforms the detection from a purely 2D image analysis problem to a 3D spatial problem, enabling the system to determine whether a detected user region represents a genuine intrusion (sufficient depth difference) or a false positive (insufficient depth difference).
2Reliability
If the system restricts door operations when intrusion is detected, then safety is improved, but false detections may cause unnecessary operational restrictions
Solution Approach 1:
The patent introduces depth difference calculation as an intermediary verification step between user region detection and door operation restriction. This intermediary mechanism filters false positives by verifying that detected users are at appropriate distances from the door, thereby improving reliability without requiring overly complex detection algorithms.
Solution Approach 2:
The system performs preliminary depth verification before executing door operation restrictions. By calculating depth differences in advance and comparing them against threshold values, the system prepares verification data beforehand, ensuring that only genuine intrusions trigger operational restrictions and reducing false alarms.
Data Source
AI summary
A door controller including an electronic control unit that controls opening and closing of a door based on imaging-data is provided. The electronic control unit is configured to recognize a user region including a user in a frame-image included in the imaging-data; extract an overlapping portion between the user region and an intrusion determination region adjacent to the openable and closable door; when a degree of overlapping of the overlapping portion relative to the intrusion determination region becomes equal to or higher than a predetermined degree, calculate a depth difference between the intrusion determination region and the overlapping portion in the frame-image; and when the depth difference is equal to or larger than a threshold value, determine that the user has intruded into the intrusion determination region.