Fisheye Image Segmentation for Free Marked Target Area Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for detecting free parking spaces using camera systems are limited by the need for accurate camera calibration and a flat environment, and struggle with variations in road markings, leading to reduced recognition quality and reliance on sparse point cloud information for free space detection.

Innovation Solution

A machine learning system is trained for semantic segmentation of images from a camera system, using fisheye lens images without rectification, to identify and segment individual free marked target areas such as parking spaces, enabling robust detection and autonomous movement towards these areas.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If classic image processing and model-based detection are used, then parking space detection can be performed, but recognition quality reduces when camera calibration is inaccurate or environment is not flat

Engineering Contradiction:

Improverecognition qualityVSAvoidrobustness to calibration errors and environmental variations

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent changes the fundamental parameters of the detection approach by transitioning from geometric model-based detection to machine learning-based detection. The system learns optimal detection parameters from training data, enabling it to adapt to various camera calibrations and environmental conditions without requiring precise manual calibration or flat environment assumptions.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mechanical/geometric calibration system with a data-driven machine learning system. Instead of relying on precise camera calibration and geometric transformations, the system uses neural networks trained on diverse images to directly detect parking spaces, eliminating the need for manual calibration and making the system robust to environmental variations.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Ease of manufacture

If fisheye lens images are used without rectification, then processing is simplified, but traditional algorithms struggle with accurate detection

Engineering Contradiction:

Improveprocessing simplicityVSAvoiddetection accuracy

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent replaces traditional geometric processing methods with machine learning-based image processing. The neural network is trained to directly process fisheye lens images without rectification, learning to compensate for the distortion automatically. This substitution enables both simplified processing and maintained detection accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the processing approach from geometric transformation-based to learning-based. Instead of rectifying the fisheye images through complex geometric transformations, the system learns optimal detection parameters directly from the distorted images, achieving both processing simplicity and detection accuracy.

Inventive Principle:

Principle #35Parameter changes

3Difficulty of detecting and measuring

If multiple algorithms are used for line detection and free space recognition, then detection capability is enhanced, but system complexity increases

Engineering Contradiction:

Improvedetection capabilityVSAvoidsystem complexity

Core Design Contradiction:

Difficulty of detecting and measuringVSDevice complexity

Solution Approach 1:

The patent merges multiple detection functions into a single machine learning model. Instead of using separate algorithms for line detection, parking space detection, and free space recognition, the neural network performs all these functions simultaneously through semantic segmentation, significantly reducing system complexity while maintaining or enhancing detection capability.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent creates a universal detection system that performs multiple functions through a single machine learning model. The semantic segmentation network simultaneously detects parking spaces, identifies road markings, recognizes free spaces, and handles various environmental conditions, replacing multiple specialized algorithms with one multi-functional system.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20240371179A1Detecting individual free marked target areas from images of a camera system of a movement device

Publication Date: 2024.11.07 CONTINENTAL AUTONOMOUS MOBILITY GERMANY GMBH

AI summary

The present disclosure relates to a method and a device for detecting individual free marked target areas such as, e.g., parking/landing/docking places, from images of a camera system) of a movement device by means of a machine learning system, and to a method for training the machine learning system. The method for segmenting and identifying individual free marked target areas includes the steps of: capturing at least one image of the environment of the movement device by means of the camera system; segmenting the individual free marked target areas from the at least one image by means of a machine learning system trained; and outputting, to a control unit, a first segment which corresponds to the individual free marked target area and second segments of at least one additional class of the captured environment of the movement device.