Road Marker-Based Object Localization for Distant Vehicle Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Infrastructure based perception systems face challenges in accurately detecting static and dynamic obstacles and determining the location of ego vehicles due to limitations in sensor range and accuracy, which can lead to unsafe or inefficient navigation.
Innovation Solution
Utilizing specialized markers on or near the road, such as solid lines, patterns, and poles, to enhance the detection and localization of objects through camera-based sensor data, enabling accurate determination of object locations and velocities using relative positioning and machine learning techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional sensor-based detection is used, then the system can operate without additional infrastructure, but the detection accuracy and location precision are insufficient
Solution Approach 1:
The patent introduces markers as intermediary objects placed on the road to mediate between the camera sensor and the objects being detected. These markers provide known reference points that enable accurate calculation of object locations, velocities, and accelerations through image processing, resolving the contradiction by improving measurement precision without requiring complex sensor systems
Solution Approach 2:
The patent creates a simplified 2D representation of the 3D environment by detecting markers and objects in camera images. This copying approach allows the system to determine accurate spatial relationships and motion parameters from 2D image data combined with known marker positions, achieving high precision without complex 3D sensing equipment
2Length of stationary object
If sensor range is extended to detect distant objects, then more of the environment can be monitored, but accuracy decreases due to distance
Solution Approach 1:
Markers serve as intermediary reference points that bridge the gap between the camera and distant objects. By detecting the relative positions of objects with respect to markers at known locations, the system can accurately determine object positions at various distances without suffering from the typical distance-accuracy tradeoff
Solution Approach 2:
The patent transitions from 3D spatial reasoning to 2D image plane analysis by working with marker and object detections in camera images. Combined with known marker positions in 3D space, this dimensional transformation enables accurate object location calculation regardless of distance, as the computation occurs in the 2D image domain where precision is maintained
3Reliability
If multiple sensors are deployed to improve detection reliability, then coverage increases, but system complexity and cost increase
Solution Approach 1:
The patent creates virtual copies of physical reference points through markers with known positions stored in a map database. These digital twins of physical markers enable the single camera to achieve reliable detection by comparing detected marker positions against the stored map information, providing redundancy without additional physical sensors
Solution Approach 2:
The system performs preliminary actions by pre-mapping marker positions and characteristics before actual object detection. This advance preparation of reference data enables the camera to reliably detect and locate objects by comparing against known marker positions, achieving detection reliability without requiring multiple simultaneous sensors
Data Source
AI summary
Methods, systems, and non-transitory computer-readable media are configured to perform operations comprising detecting an object in sensor data captured in an environment; detecting a marker in the sensor data captured in the environment; and determining a first location of the object in the environment based on a relative location of the object to the marker and a second location of the marker.