Lidar Signal Processor Waveform Classification for Overlapping Targets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional lidar apparatuses struggle to accurately distinguish between closely overlapping targets and obliquely positioned targets using only specific points in the received waveform, leading to difficulties in identifying and separating overlapped waveforms.
Innovation Solution
A lidar apparatus with a signal processor that generates information including coordinates of points on rising and falling edges, classifies pixels based on peak start and end points, and performs additional classification using neighboring pixel information to determine overlapping or non-overlapping types, allowing for the separation of overlapped waveforms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If only specific points in the received waveform are used as output, then data transmission efficiency is improved, but the ability to distinguish between closely overlapping targets and obliquely positioned targets deteriorates
Solution Approach 1:
The patent segments the received waveform analysis into multiple classification stages: first classifying pixels based on peak characteristics (single peak vs. multiple peaks), then further classifying ambiguous cases using neighboring pixel information. This segmented approach enables accurate target distinction while maintaining efficient data transmission by only processing necessary points.
Solution Approach 2:
The patent introduces a new dimension of analysis by comparing peak positions across multiple pixels (spatial dimension) in addition to temporal waveform characteristics. By examining whether peak positions coincide across neighboring pixels, the system can distinguish overlapping targets from oblique targets, resolving the accuracy-efficiency contradiction.
2Measurement precision
If all received waveforms within the field of view are output as they are, then measurement completeness is improved, but data transmission burden increases making real-time transmission difficult
Solution Approach 1:
The patent extracts only the essential characteristic points from complete waveforms for transmission: peak positions, peak intensities, and pixel coordinates. By taking out only these critical features rather than transmitting entire waveforms, the system maintains measurement completeness while enabling real-time data transmission with significantly reduced data volume.
3Device complexity
If conventional lidar apparatuses use only specific points to represent targets, then device complexity is reduced, but the ability to separate overlapped waveforms deteriorates
Solution Approach 1:
The patent performs preliminary classification of pixels based on peak characteristics before final target identification. By预先 categorizing pixels into single-peak and multi-peak groups and identifying ambiguous cases, the system prepares structured information that simplifies subsequent waveform separation while maintaining high separation accuracy for overlapped targets.
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
Enables accurate identification and separation of overlapped waveforms, effectively distinguishing between closely overlapping and obliquely positioned targets, improving the efficiency of signal processing and real-time output in lidar systems.
Implementation Method 1
an optical transmitter configured to transmit a laser light for external detection
Implementation Method 2
an optical receiver configured to receive the laser light reflected from the external
Data Source
AI summary
A lidar apparatus is disclosed. The lidar apparatus according to an embodiment of the present disclosure includes an optical transmitter configured to transmit a laser light for external detection; an optical receiver configured to receive the laser light reflected from the external; and a signal processor configured to detect the laser light received by the optical receiver, wherein the signal processor may be configured to: generate information including coordinates of a plurality of points including at least one point on a rising edge, a start point of a peak, an end point of the peak, and at least one point on a falling edge of a received waveform of the laser light reflected from the external arbitrary area as pixel information about a pixel corresponding to the arbitrary area, and classify the pixel according to whether the start point of the peak and the end point of the peak are the same.