Periodic Stationary Object Detection via Bird's Eye View Waveform Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing object detection systems have difficulty distinguishing and recognizing periodically present stationary objects like pylons, guardrail posts, or utility poles from other three-dimensional objects.

Innovation Solution

A periodic stationary object detection system that includes an image-capturing device, a viewpoint conversion unit, a feature point-extracting unit, a waveform data-calculating unit, a peak information detecting unit, a periodic stationary object candidate-detecting unit, and a periodic stationary object-judging unit to identify periodic stationary objects by converting images to bird's eye view, extracting feature points, calculating waveform data, and determining peak information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a simple object detection system is used, then the device complexity is low, but the measurement precision for periodic stationary objects deteriorates

Engineering Contradiction:
Improvedetection precision of periodic stationary objectsVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The detection system segments the detection process into distinct functional modules: viewpoint conversion unit, feature point extraction unit, waveform data calculation unit, peak information detection unit, and periodic stationary object judging unit. Each module handles a specific aspect of the detection task, allowing the system to achieve high precision detection of periodic stationary objects while maintaining manageable complexity through modular design.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs viewpoint conversion to create a bird's-eye view image from the captured image. This dimensional transformation allows the system to analyze the spatial distribution of feature points in a top-down perspective, making it easier to identify periodic patterns of stationary objects along the roadside by examining their positional relationships in the converted coordinate system.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If viewpoint conversion and waveform analysis are performed, then the detection precision improves, but the loss of time increases

Engineering Contradiction:
Improvedetection precisionVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs viewpoint conversion on the captured image to create a bird's-eye view image before extracting feature points. This preliminary transformation establishes a consistent reference frame that simplifies subsequent processing steps, allowing feature points to be extracted more efficiently and reducing the overall processing time despite the additional initial computation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The feature point extraction unit selectively extracts only the relevant feature points from the bird's-eye view image that correspond to potential stationary objects. By focusing computation on extracting and analyzing only these specific feature points rather than processing the entire image, the system achieves high detection precision while minimizing unnecessary computational overhead and processing time.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP2680247B1Periodic stationary object detection device and periodic stationary object detection method
Publication Date: 2018.12.26 NISSAN MOTOR CO LTD
  • EP2680247B1 patent drawingFigure 1~2
  • EP2680247B1 patent drawingFigure 3
  • EP2680247B1 patent drawingFigure 4(a)~4(b)

AI summary

A periodic stationary object detection system extracts a feature point of a three-dimensional object from image data on a predetermined region of a bird's eye view image for each of multiple sub regions included in the predetermined region, calculates waveform data corresponding to a distribution of the feature points in the predetermined region on the bird's eye view image, and judges whether or not the three-dimensional object having the extracted feature point is a periodic stationary object candidate on the basis of whether or not peak information of the waveform data is equal to or larger than a predetermined threshold value.