Motor Vehicle Object Distance Estimation Using Partial Visibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for estimating the distance of an object from a motor vehicle using an image sensor fail when the object exceeds the sensor's field of view, particularly at low relative speeds, leading to delayed warnings or braking activations.
Innovation Solution
A method that determines a first value of the distance parameter based on the object's dimensions in a fully visible image and calculates a second value using the dimensions of the object's partially visible portion, employing geometric shapes and visual tracking algorithms to maintain estimation even when the object is not fully contained within the image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the evaluation is based on the dimensions of the object in the image, then the parameter can be determined when the object is fully visible, but the evaluation fails when the object exceeds the useful field of the image sensor
Solution Approach 1:
The patent segments the object recognition task into two parts: when the object is fully visible, use its complete dimensions for parameter estimation; when the object is partially visible, use only the visible portion's dimensions. This segmentation allows the system to adapt its measurement approach based on object visibility, maintaining reliability across different scenarios.
Solution Approach 2:
The patent applies partial action by using only the visible portion of the object when it exceeds the image sensor's field of view. Instead of requiring the complete object to be visible, the system successfully estimates parameters using merely the visible part, thereby maintaining functionality under partial visibility conditions.
2Productivity
If the object is partially outside the image sensor's field of view, then the system can continue to estimate the parameter using the visible portion, but the measurement precision may be reduced
Solution Approach 1:
The patent performs preliminary action by detecting and tracking the object in advance when it is fully visible within the field of view. This early detection establishes baseline measurements and object characteristics that can be used to maintain estimation accuracy even when the object later moves partially outside the field of view.
Solution Approach 2:
The system uses feedback from continuous tracking of the visible object portion to update and maintain parameter estimates. By continuously monitoring the visible part of the object and comparing it against expected object models, the system can compensate for the reduced information available when the object is partially outside the field of view.
Data Source
Figure 1~3
Figure 4
AI summary
The invention relates to a method for estimating a parameter representing the distance of an object (4), by means of an image sensor provided on a motor vehicle. Said method includes the following steps: determining a first value of the parameter on the basis of a size of the object in at least one first image (20, 22), taken by the image sensor and in which the object is entirely contained; determining a size of only one portion (44, 46) of the object in a second image (24, 26), taken by the image sensor and in which the object is only partially contained; and determining a second value of the parameter on the basis of the size of said portion (44, 46).