Moving-Object Detection Using Odometry and Image-Point Cloud Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Detecting moving objects in crowded environments is challenging due to occlusions and complex interactions, especially when the source of visual information is moving, which complicates motion prediction for autonomous vehicles.
Innovation Solution
A method and apparatus that utilize a machine-learning model trained with odometry data and visual data from image and point-cloud data to detect moving objects by leveraging motion consistency and magnitude-of-motion comparisons, reducing computational resources and memory requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional moving-object detection methods are used in crowded environments, then detection capability is limited, but computational resources and memory requirements increase significantly
Solution Approach 1:
The system segments the detection task by separating static and moving objects through odometry-based motion compensation. By compensating for sensor motion and isolating actual moving objects from the static scene, the system reduces the search space and computational burden while improving detection accuracy in crowded environments.
Solution Approach 2:
Odometry data serves as an intermediary that bridges the gap between sensor motion and scene understanding. The odometry information is used to compensate for sensor movement and generate motion masks, enabling the system to distinguish between apparent motion caused by sensor movement and actual object motion, thereby reducing false positives and computational overhead.
2Measurement precision
If detailed processing of image and point-cloud data is performed to improve detection accuracy, then measurement precision improves, but device complexity increases
Solution Approach 1:
The system performs preliminary motion compensation using odometry data before executing the main detection algorithm. By pre-compensating for sensor motion and generating motion masks in advance, the system simplifies the subsequent detection process and reduces the complexity of the overall system while maintaining high detection precision.
Solution Approach 2:
The system integrates data from multiple dimensions including image data, point-cloud data, and odometry data. By fusing information across these different dimensional sources and processing them in a unified framework, the system achieves high detection precision without requiring overly complex processing of individual data types.
3Reliability
If motion prediction is performed for autonomous vehicle navigation, then navigation safety improves, but loss of time increases due to computational requirements
Solution Approach 1:
The system extracts only the essential motion information needed for navigation safety by using odometry-based motion masks to isolate moving objects. By extracting and focusing on relevant motion patterns rather than processing all scene details, the system enables timely motion prediction while maintaining navigation safety.
Solution Approach 2:
The odometry system serves itself by providing motion compensation information that directly enables the detection and prediction pipelines. The self-contained nature of the odometry-based approach allows the system to perform motion prediction efficiently without requiring external computational resources, reducing time loss while maintaining reliability.
Data Source
AI summary
Systems and techniques are described herein for detecting objects. For instance, a method for detecting objects is provided. The method may include obtaining image data representative of a scene and point-cloud data representative of the scene; processing the image data and the point-cloud data using a machine-learning model, wherein the machine-learning model is trained using at least one loss function to detect moving objects represented by image data and point-cloud data, the at least one loss function being based on odometry data and at least one of training image-data features or training point-cloud-data features; and obtaining, from the machine-learning model, indications of one or more objects that are moving in the scene.