Dynamic Vision Sensor Directing Lidar Scanning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lidar systems face limitations in frame rate due to increased distance and detail, leading to higher latency and motion artifacts, and existing methods for selecting areas of interest are slow and computationally expensive, which can result in missed important objects in fast-changing scenes.
Innovation Solution
A Dynamic Vision Sensor (DVS) camera is used to detect illumination changes with low latency, directly controlling the lidar to scan areas of interest without the need for frame capture or complex object detection, allowing for reduced latency and higher resolution scans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the lidar scans more rows and points to increase frame size and resolution, then the measurement precision and detail detection improve, but the frame rate decreases and latency increases
Solution Approach 1:
The patent divides the full frame scan into multiple independent row segments that can be scanned and processed separately. The controller can selectively enable or disable specific rows based on event data, transforming a monolithic scanning process into modular segments that can be dynamically adjusted to balance resolution and frame rate.
Solution Approach 2:
The patent implements dynamic row enabling/disabling based on real-time event data from the DVS camera. The controller dynamically adjusts which rows are actively scanned by modifying the scanner control signals, allowing the system to adapt the scanning pattern to current scene changes and optimize between resolution and frame rate in real-time.
2Area of stationary object
If the lidar scans more rows to increase frame size, then the coverage area improves, but the scanning time increases and frame rate decreases
Solution Approach 1:
The patent extracts and processes only the necessary portions of the scene by selectively enabling specific rows based on DVS event data. Instead of scanning the entire coverage area uniformly, the system extracts and focuses scanning resources on regions with detected changes, reducing unnecessary scanning time while maintaining adequate coverage.
Solution Approach 2:
The patent implements partial scanning by enabling only a subset of rows that contain relevant information based on event detection. This partial action approach scans fewer rows than a full frame scan would require, thereby reducing scanning time and latency while still capturing the essential changes in the scene.
3Loss of information
If a visual light camera is used to select areas of interest, then the lidar can focus on important regions, but the processing time increases and important objects may be missed in fast-changing scenes
Solution Approach 1:
The patent replaces the traditional frame-based image processing system with an event-driven detection system. Instead of capturing and processing complete video frames to identify areas of interest, the DVS camera generates asynchronous event signals directly from detected changes, eliminating the computational overhead of frame processing and enabling real-time response to scene changes.
Solution Approach 2:
The patent implements continuous monitoring through the DVS camera's event-driven architecture, which continuously detects illumination changes without the间断s inherent in frame-based systems. This continuous useful action ensures that important objects in fast-changing scenes are detected immediately as changes occur, rather than being missed between frames.
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
The DVS camera enables rapid identification and tracking of moving objects, allowing for quicker and more accurate updates to the 3D model, reducing the risk of missing important objects and improving overall lidar performance in dynamic environments.
Implementation Method 1
A dynamic vision sensor (DVS) camera is used to detect illumination changes
Implementation Method 2
lidar systems scan the environment to build a point cloud
Implementation Method 3
the light pulse's travel time from the lidar to an object and back
Data Source
AI summary
A dynamic vision sensor (DVS) camera is described that directs scanning of a lidar. In one example, a DVS has an array of pixels, wherein each DVS pixel comprises a photodetector, and is configured to detect temporal contrast and generate an event in response. A lidar system has a light source configured to emit light, a scanner configured to direct the emitted light along a scan pattern contained within a field of regard of the lidar system, and a receiver configured to detect at least a portion of the emitted light scattered by one or more remote targets in a scene. A processor is coupled to the lidar system and to the dynamic vision sensor and receives the events, identifies a region of interest in the field of regard that corresponds to DVS pixels that generated the events, and adjusts a scan parameter of the lidar system in the region of interest.