3D Point Cloud Edge Pixel Filtering for ToF Depth Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional edge detection methods for time-of-flight sensors in autonomous vehicles are inefficient and computationally expensive, leading to inaccurate depth information due to mixed reflections from foreground and background objects, which can result in false positives in object identification and tracking.
Innovation Solution
Systems and techniques for identifying and filtering edge pixels by deriving information in the phase and/or distance domain, using a depth frame differential operation to create an edge pixel distribution map, and applying adaptive or constant thresholds for accurate labeling and filtering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional edge detection methods are used for time-of-flight sensors, then object detection can be performed, but computational cost and processing time increase significantly
Solution Approach 1:
The patent segments the depth map processing into distinct stages: obtaining a first depth map, generating a second depth map through shifting operations, comparing the two maps to identify edge pixels, and finally filtering based on edge pixel distribution. This segmentation allows efficient processing by breaking down the complex edge detection task into manageable steps that can be executed sequentially with optimized computational resources.
Solution Approach 2:
The patent performs preliminary actions by creating multiple shifted versions of the depth map before the actual edge detection occurs. By pre-generating these shifted depth maps and identifying potential edge regions in advance, the system reduces the computational burden during the final edge detection and filtering stage, thereby decreasing overall processing time while maintaining detection accuracy.
2Measurement precision
If conventional edge detection methods are used for time-of-flight sensors, then edge pixels can be identified, but computational burden and power consumption increase
Solution Approach 1:
The patent extracts only the essential information needed for edge pixel identification by comparing shifted depth maps and focusing specifically on regions where depth differences exceed a threshold. This extraction approach avoids unnecessary computational operations on non-edge regions, reducing power consumption while maintaining accurate edge pixel identification. The system takes out only the relevant edge pixel data from the entire depth map for further processing.
Solution Approach 2:
The patent applies partial action by performing depth map shifting and comparison operations only on regions where edges are likely to occur, rather than processing the entire depth map uniformly. By using adaptive thresholds that focus computational effort on critical regions, the system achieves accurate edge detection with reduced overall computational burden and lower power consumption compared to exhaustive conventional methods.
3Ease of manufacture
If mixed reflections from foreground and background objects are not filtered, then processing is simpler, but depth information accuracy decreases
Solution Approach 1:
The patent introduces an intermediary filtering mechanism that processes depth information between acquisition and final object detection. The edge pixel identification and filtering system acts as a mediator that separates mixed reflections from foreground and background objects by identifying and excluding edge pixels that contain contaminated depth data. This intermediary step maintains processing simplicity while significantly improving depth information accuracy through targeted filtering.
Solution Approach 2:
The patent converts the harmful effect of mixed reflections into a beneficial filtering opportunity. By detecting edge pixels where mixed reflections occur (through depth map comparison and thresholding), the system identifies precisely which data points need filtering. The presence of mixed reflections actually helps locate edge regions, which then become the focus of the filtering operation, ultimately improving overall depth map quality by removing contaminated data.
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
Improve the accuracy of depth information by effectively distinguishing edge pixels, reducing computational burden and power consumption, thereby enhancing the reliability of object detection and tracking in autonomous vehicles.
Implementation Method 1
depth information from a time-of-flight (ToF) sensor
Data Source
AI summary
Systems and techniques are provided for identifying and filtering edge pixels from a 3D point cloud from a time-of-flight sensor. An example method includes receiving a first depth map that is based on an image frame captured by a time-of-flight (ToF) sensor, wherein the first depth map includes a plurality of measurements corresponding to a pixel array of the ToF sensor; generating a second depth map by shifting the plurality of measurements corresponding to the pixel array in at least one direction; comparing the first depth map with the second depth map to determine a measurement difference for each pixel in the pixel array; and identifying one or more edge pixels in the pixel array corresponding to at least one edge region in the image frame, wherein the measurement difference associated with the one or more edge pixels is greater than an edge threshold.


