Monocular Camera Signal Analysis for Drawing Object Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Automated vehicles equipped with monocular cameras face challenges in distinguishing between real roads and drawing objects, such as photographs or pictures, leading to potential collisions when erroneous recognition occurs.

Innovation Solution

A signal processing device and method that analyzes images captured by a monocular camera to determine whether an object is a real object or a drawing object by evaluating changes in the focus of expansion (FOE) position, lane width, and ground level position, using thresholds to differentiate between real and drawn features.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a monocular camera is used for object detection, then the device complexity is reduced, but the measurement precision of object distance and real vs. drawing object discrimination deteriorates

Engineering Contradiction:
Improvecamera system complexityVSAvoidobject distance measurement precision
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The system performs preliminary analysis of the captured image to detect candidate drawing objects before final object recognition. By pre-identifying regions that contain drawing characteristics (such as photographs or pictures attached to walls), the system可以避免误识别这些绘图对象为真实道路,从而解决了单目相机深度测量精度不足导致的问题

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediate processing stage that analyzes image characteristics to distinguish drawing objects from real objects. This intermediary analysis layer examines features such as texture patterns, edge characteristics, and contextual clues to determine whether detected objects are real or drawn, thereby compensating for the monocular camera's inability to accurately measure distance

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If image analysis is performed to distinguish drawing objects from real objects, then the reliability of automated driving is improved, but the processing time increases

Engineering Contradiction:
Improveautomated driving reliabilityVSAvoidimage processing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The image analysis process is segmented into multiple stages: first detecting candidate drawing objects, then analyzing their characteristics, and finally making recognition decisions. This segmentation allows the system to focus computational resources only on suspicious regions rather than processing the entire image, thereby reducing overall processing time while maintaining high reliability

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs partial analysis by focusing only on regions that exhibit drawing characteristics rather than analyzing the entire image in detail. By applying simplified detection algorithms to identify candidate regions and then performing more thorough analysis only on these specific areas, the system achieves high reliability without excessive processing time

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS20240371173A1Signal processing device, signal processing method, and program
Publication Date: 2024.11.07 SONY SEMICON SOLUTIONS CORP
  • US20240371173A1 patent drawing
  • US20240371173A1 patent drawing
  • US20240371173A1 patent drawing

AI summary

To realize a configuration of analyzing a captured image captured by a monocular camera and determining whether an object in the captured image is a real object or a drawing object. An image captured by a monocular camera mounted on a vehicle is analyzed to determine whether an object in the captured image is a real object or a drawing object. An image signal analysis unit determines that an object in the captured image is a drawing object in a case where a change amount per unit time of a FOE (focus of expansion) position in the captured image is equal to or larger than a predetermined threshold value, in a case where a change amount per unit time of a lane width detected from the captured image is equal to or larger than a predetermined threshold value, or in a case where a difference between a grounding position of a vehicle or the like in the captured image and a ground level position corresponding to the vehicle grounding position is equal to or larger than a predetermined threshold value.