Computer Vision Object Detection for Lightweight Semantic Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mapping and navigation services face challenges in extracting useful and lightweight information from real-time sensor data and inferred scene data for location-based services, particularly in supporting intelligent transportation systems, due to the high precision and large data volumes required for advanced driving assistance systems.
Innovation Solution
A computer vision-based method that processes images from vehicles or devices traveling at street level to detect objects, lane markings, and road surfaces, determining their relative positioning and classifying semantic localization features, which are then provided as output, enabling efficient object detection and localization without requiring precise location data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-precision object detection is used for advanced driving assistance systems, then detection accuracy is improved, but data volume and processing complexity increase tremendously
Solution Approach 1:
The patent extracts only the essential semantic localization features (relative position, size, shape) from complete object detection data. Instead of processing all detected object attributes, the system selectively extracts only the features necessary for location-based services, thereby reducing data volume while maintaining detection precision for the extracted features.
Solution Approach 2:
The patent segments the object detection process into two stages: first, standard object detection for accuracy; second, selective extraction of semantic localization features. This segmentation allows the system to maintain high detection precision in the first stage while reducing data processing burden in the second stage by only extracting relevant features.
2Measurement precision
If high-precision object detection is used for advanced driving assistance systems, then detection accuracy is improved, but processing complexity increases tremendously
Solution Approach 1:
The patent extracts only the essential semantic localization features (relative position, size, shape) from complete object detection data. Instead of processing all detected object attributes, the system selectively extracts only the features necessary for location-based services, thereby reducing data volume while maintaining detection precision for the extracted features.
Solution Approach 2:
The patent segments the object detection process into two stages: first, standard object detection for accuracy; second, selective extraction of semantic localization features. This segmentation allows the system to maintain high detection precision in the first stage while reducing data processing burden in the second stage by only extracting relevant features.
3Measurement precision
If precise location data is collected for all detected objects, then location accuracy is improved, but data processing requirements increase
Solution Approach 1:
The patent applies local quality by providing different levels of location data precision based on the specific application need. For location-based services, only relative semantic localization features are extracted and provided, rather than complete precise coordinates for all objects. This allows the system to maintain high location accuracy where needed while improving overall data processing efficiency by reducing the volume of precise location data generated.
Data Source
AI summary
An approach is provided for computer-vision-based object detection. The approach involves, for example, receiving an image captured from a perspective of a vehicle or a device traveling at street level. The approach also involves processing the image using computer vision to detect one or more objects, one or more lane markings, a road surface, or a combination thereof depicted in the image. The approach further involves determining a relative positioning of the one or more objects with respect to the one or more lane markings, the road surface, or a combination thereof. The approach further involves classifying one or more semantic localization features of the one or more objects based on the relative positioning. The approach further involves providing the one or more semantic localization features as an output.