Vehicle Warning Control for Monocular Camera Distance Misjudgment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle control devices using monocular cameras for estimating distances between vehicles and objects may inaccurately estimate distances, particularly when objects are moving towards the vehicle, such as pedestrians descending stairs, leading to potential safety hazards.
Innovation Solution
A vehicle control device equipped with a surrounding sensor including a distance measuring sensor and a monocular camera, along with a processor that calculates distances using both sensor data and image analysis. The processor prioritizes warnings based on the feature amount, which is the change in the image position of a moving object relative to its size, to enhance safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If distance is estimated solely based on vertical position in monocular camera image, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent combines data from multiple sensors (monocular camera, distance measuring sensor, and potentially other surrounding sensors) to estimate distance. The processor integrates image data with depth information from the distance measuring sensor, merging multiple data sources to achieve more accurate distance estimation while maintaining reasonable system complexity.
Solution Approach 2:
The patent introduces a processor as an intermediary that receives data from both the monocular camera and distance measuring sensor, processes this information, and generates corrected distance estimates. The processor acts as a mediator that reconciles the limitations of individual sensors by combining their outputs through image processing and coordinate transformation.
2Device complexity
If monocular camera is used for distance estimation, then device complexity is reduced, but reliability deteriorates for moving objects on slopes
Solution Approach 1:
The system uses feedback from the distance measuring sensor to correct and verify distance estimates derived from the monocular camera. The processor continuously compares camera-based estimates with sensor-based measurements and adjusts the estimation algorithm accordingly, creating a feedback loop that improves reliability for moving objects including those on slopes or stairs.
Solution Approach 2:
The system performs preliminary distance measurement using the distance measuring sensor before final distance estimation is made. This preliminary action provides reference data that helps the processor anticipate and correct for scenarios involving moving objects on slopes, stairs, or uneven terrain before misestimation occurs.
3Loss of time
If warning priority is based solely on distance threshold, then response time is reduced, but object-generated harmful factors increase due to misjudgment
Solution Approach 1:
The patent changes the parameters used for warning priority determination from simple distance thresholds to a multi-parameter assessment including vertical position in image, distance measuring sensor data, and calculated feature amounts. This parameter change allows the system to maintain rapid response while reducing misjudgments by considering additional dimensions of object characteristics and position.
Solution Approach 2:
The system adds vertical dimension analysis to the warning priority determination by examining the vertical position of objects in the camera image and the change in this position over time. This dimensional addition helps distinguish between objects at different heights (such as pedestrians on stairs versus distant objects), reducing false warnings and improving safety without sacrificing response time.
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 solution improves the accuracy of distance estimation and prioritizes alerts for moving objects, thereby enhancing safety by reducing the likelihood of misjudging distances and improving response times to potential hazards.
Implementation Method 1
a surrounding sensor (20) that includes a distance measuring sensor (21) for acquiring the distance between an object located in a predetermined first area in the traveling direction of the own vehicle (V) and the own vehicle
Implementation Method 2
a monocular camera (22) for capturing a predetermined second area in the traveling direction of the own vehicle
Data Source
AI summary
The processor of the vehicle control device 1 acquires a second distance between the own vehicle and a moving object based on the position of the image of the moving object in the vertical direction within the imaging range of the monocular camera. When the second distance is greater than the first distance between the own vehicle and a stationary object, and the difference exceeds a threshold, and when the feature quantity, which is the amount of change in the position of the moving object's image relative to the amount of change in the size of the moving object's image as it moves downward and expands within the imaging range of the monocular camera, exceeds a predetermined value, the processor increases the priority of the warning process to alert attention to the moving object.