Sensor Target Labeling With Camera-Distance Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for automatic object and segment labeling of sensor target data from vehicle sensors, such as cameras and lidar, struggle with accuracy and reliability, particularly in distinguishing between static and dynamic objects and segments, due to limited fusion of camera-based and distance-based data.
Innovation Solution
A method that combines camera images with distance sensor data, using learned machine recognition methods to generate an environment representation, adjust it with distance data, and calculate a synthetic image for accurate labeling, incorporating depth and reflection properties to enhance object and segment classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If camera images are used alone for object labeling, then the system complexity is low, but the labeling accuracy and reliability deteriorate
Solution Approach 1:
The patent combines camera images with distance sensor data (lidar, radar, ultrasonic sensors) to create a fused sensor input system. This merging of multiple sensor types enables more accurate object detection and labeling by leveraging both visual appearance and spatial distance information, directly resolving the contradiction between labeling accuracy and system complexity
Solution Approach 2:
The system processes multiple sensor types (camera, lidar, radar, ultrasonic) through a unified machine recognition method that handles diverse data formats. This multi-functional approach allows the same processing pipeline to work with different sensor modalities, improving labeling accuracy while managing system complexity through standardized processing
2Reliability
If multiple sensor types are integrated for data fusion, then the labeling reliability improves, but the processing complexity increases
Solution Approach 1:
The patent implements sensor fusion by combining data from camera, lidar, radar, and ultrasonic sensors into a unified processing framework. This merging approach improves labeling reliability by cross-validating detections across multiple sensor types while managing processing complexity through integrated algorithms
Solution Approach 2:
The system uses machine recognition methods that process fused sensor data and generate labeled outputs that can be used for training and validation. This feedback mechanism improves reliability by continuously refining the labeling accuracy based on processed results while managing complexity through iterative optimization
3Measurement precision
If distance sensor data is captured and processed, then the depth information accuracy improves, but the energy consumption increases
Solution Approach 1:
The system captures distance data from multiple sensor types (lidar, radar, ultrasonic) but processes this data selectively through machine recognition methods. By applying partial processing to the fused data rather than fully processing all sensor inputs, the system achieves improved depth accuracy while managing energy consumption through optimized processing thresholds
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
This approach improves the accuracy and reliability of sensor target data labeling by integrating multiple sensor types, providing high-quality training data for machine recognition methods, and enhancing user understanding of the vehicle environment.
Implementation Method 1
a distance between the vehicle and an object point and/or a position of the object point of the object in the environment of the vehicle can be ascertained from the transit time of the reflected signal
Implementation Method 2
the distance sensor advantageously receives the signal reflected on objects in the environment of the vehicle and related to the sent transmission signal
Data Source
AI summary
A method for automatic object and/or segment labeling of sensor target data of at least one vehicle target sensor. The method comprises first capturing of at least one sequence of camera images; generating an environment representation of the vehicle as a function of the captured sequence; recognizing at least one object in the environment by a learned machine recognition method as a function of a captured camera image; ascertaining an estimated position of the object as a function of the camera image; classifying a point of the environment representation based on the recognized object and the ascertained estimated position; and second capturing of distance data using at least one distance sensor. The generated environment representation is adjusted as a function of the captured distance data. A calculation of a synthetic image of the environment from a virtual perspective of observation takes place based on the adjusted environment representation.