Test Device for a Motor Vehicle Navigation Device and Method for Motion Detection in an Image Sequence

Simple SummaryContent extracted from patent full text and abstract with AI.

This invention describes a system and method for robustly detecting and encoding movements in image sequences, specifically tailored for test devices assessing automotive navigation systems. The core innovation is restricting motion detection and feature matching to dynamic (moving) regions of the display—such as animated map segments—by applying an adaptive mask that excludes static areas (like menus or status bars). This targeted approach improves the accuracy and efficiency of movement recognition and video compression, even when both static and moving objects are visible in the images.

Use CasesContent extracted from patent full text and abstract with AI.

• Testing and evaluating automotive navigation systems by recording and analyzing their on-screen map and navigation displays during test drives.
• Improving video encoding efficiency for in-vehicle navigation system recordings by compressing dynamic map areas more effectively.
• Robust motion tracking in any application where image sequences show both static and dynamic elements (e.g., surveillance systems showing moving vehicles on a fixed background).
• Automated quality assurance systems for software that displays rich, dynamic content alongside static UI components.
• Real-time video analysis in transport-related applications to distinguish between static information and relevant moving map content.

BenefitsContent extracted from patent full text and abstract with AI.

• Enables highly accurate and robust movement detection in scenes with both moving and static elements.
• Enhances data compression, allowing longer recordings without needing increased memory, by focusing encoding on dynamic regions.
• Maintains compatibility with efficient, real-time-capable feature detection algorithms (like ORB), even in the presence of static text-heavy UI elements.
• Yields superior model parameter estimation for global motion compensation, improving video encoding quality and efficiency.
• Reduces processing and storage demands by ignoring irrelevant, unchanging image areas during feature extraction and motion analysis.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

The invention relates to a method for recognising movements in an image sequence containing a current image (A) and a reference image (Ref1), wherein the following steps are executed by an image processing device: detecting features in the current image (A) and in the reference image (Ref1); and forming pairs of features (M1-M1' to M4-M4') by combining respective image features that belong together (M1 to M4, M1' to M4'). The invention addresses the problem of enabling robust recognition of movements in an image sequence in which images of static and moving objects are shown simultaneously. For this purpose, features are detected in the images (A, Ref1) only in a predetermined search area and the search area is established using a mask by masking the images (A, Ref1).