Object Tracking Across Local and Global Maps with Kalman Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection and classification systems face challenges in achieving robust, real-time detection and localization of a broad range of objects across various settings, with issues such as false detections, duplicate detections, and localization errors, particularly in systems with memory, processing, and resource constraints.
Innovation Solution
A system utilizing a multi-modal, multi-camera unmanned vehicle equipped with sensors like visible light, infrared, radar, and Lidar, combined with a first logic device that fuses sensor inputs using a trained neural network, Kalman filtering, and object tracking modules to detect, classify, and localize objects within local and global maps, allowing for user-in-the-loop corrections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a trained neural network is used for object detection and classification, then detection accuracy is improved, but system complexity and resource requirements increase
Solution Approach 1:
The system segments the object detection task into multiple specialized neural networks, each trained for specific object classes or detection scenarios. This allows the system to achieve high accuracy for specific tasks while keeping individual network complexity manageable, and enables selective deployment based on operational needs.
Solution Approach 2:
The system implements a universal object detection framework that can handle multiple object classes, detection modes, and operational scenarios through a common architecture. The neural network system is designed to be adaptable to various end-use scenarios while maintaining a unified processing pipeline, reducing overall system complexity.
2Reliability
If multiple sensors and cameras are used for comprehensive object detection, then detection reliability is improved, but processing complexity and computational load increase
Solution Approach 1:
The system merges data from multiple sensors and cameras into a unified detection framework. By combining sensor inputs at the processing level rather than maintaining separate processing pipelines, the system achieves improved detection reliability through multi-sensor fusion while reducing overall processing complexity through integrated architecture.
Solution Approach 2:
The system introduces an intermediary processing layer that standardizes and harmonizes data from different sensor types before feeding them to the neural networks. This mediator layer handles sensor calibration, coordinate transformation, and data synchronization, making multi-sensor integration more manageable and reducing computational overhead.
3Measurement precision
If real-time object detection is implemented with comprehensive training data, then detection accuracy across various settings is improved, but training time and data collection costs increase
Solution Approach 1:
The system performs preliminary actions by pre-training neural networks on comprehensive datasets covering various object classes and scenarios. These pre-trained networks serve as foundational models that can be quickly adapted to specific operational scenarios through fine-tuning, significantly reducing the time required for full training while maintaining high detection accuracy across diverse settings.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A detection device, such as an unmanned vehicle, is adapted to traverse a search area and generate sensor data associated with objects that may be present in the search area. The generated sensor data is used by a system including object detection inference models configured to receive the sensor data and output object data, a local object tracker configured to track detected objects in a local map, and a global object tracker configured to track detected objects on a global map. The local object tracker is configured to fuse object detections from the object detection inference models to identify locally tracked objects, and a Kalman filter processes frames of fused object data to resolve duplicates and/or invalid object detections. The global object tracker includes a pose manager, configured to track global objects in the global map and update the pose based on a map optimization process. User-in-the-loop processing includes a user interface for displaying and manual editing of detected object data.