Dynamic Privacy Mask Generation for PTZ Camera Field of View Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems for PTZ cameras face challenges in generating appropriate privacy masks when the field of view changes, leading to exposure of privacy areas and increased computational complexity.
Innovation Solution
An image processing method that converts X-Y plane coordinates of a rectangular mask's endpoints into spherical coordinates, allowing for the generation of a new mask in a changed field of view by determining the maximum and minimum X-Y coordinates, and adjusting the mask boundaries to ensure the privacy area remains within the image frame, even if partially out of bounds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the position and/or size of the privacy area changes based on user or system control for a PTZ camera, then the privacy mask can adapt to different fields of view, but the privacy area may be exposed in the image if the mask does not change correspondingly
Solution Approach 1:
The patent implements a dynamic privacy mask system that automatically adjusts the mask position and size based on real-time camera parameters (pan, tilt, zoom values). The mask coordinates are dynamically recalculated using transformation formulas that incorporate camera rotation angles and zoom levels, ensuring the mask continuously adapts to changing fields of view while maintaining reliable privacy protection.
Solution Approach 2:
The system employs feedback mechanisms by continuously monitoring camera parameter changes and using these changes to update the privacy mask coordinates. The transformation process uses feedback from camera state (pan, tilt, zoom values) to adjust mask parameters, creating a closed-loop system that ensures the mask remains aligned with the intended privacy area despite camera movements.
2Device complexity
If the position and/or size of the privacy mask does not change correspondingly with field of view changes, then the system maintains simplicity, but the privacy area will be exposed in the image
Solution Approach 1:
The patent introduces coordinate transformation formulas as intermediaries that bridge the camera parameter space and mask coordinate space. These transformation equations serve as mathematical mediators that automatically convert camera state (pan, tilt, zoom) into corresponding mask coordinates, providing a systematic yet relatively simple method to maintain privacy protection without requiring complex manual adjustments.
Solution Approach 2:
The system changes mask parameters (position and size) in response to camera parameter changes. By establishing a direct relationship between camera parameters (pan angle, tilt angle, zoom level) and mask coordinates through transformation formulas, the system automatically adjusts mask parameters when camera parameters change, maintaining privacy protection while avoiding excessive system complexity.
3Reliability
If an algorithm for generating the privacy mask becomes complicated to handle parts of the privacy area exceeding the image, then the privacy area can be fully protected, but the computation load increases
Solution Approach 1:
The patent extracts only the necessary computation by calculating mask coordinates only for the visible portion of the privacy area within the current field of view. Rather than computing complex clipping algorithms for out-of-bounds areas, the system extracts and processes only the relevant coordinate transformations for visible regions, reducing computation load while maintaining effective privacy protection for the visible privacy area.
Solution Approach 2:
The system applies partial action by generating mask coordinates only for the portion of the privacy area that is currently visible in the image. When parts of the privacy area exceed the image boundaries, the algorithm computes masks only for the in-bounds portions, avoiding unnecessary computations for out-of-bounds areas while still providing complete protection for all visible privacy regions.
Data Source
AI summary
An image processing method includes steps of obtaining a first image under a first field of view; generating a first rectangular mask in the first image; converting first X-Y plane coordinates of four end points of the first rectangular mask into spherical coordinates; obtaining a second image under a second field of view; converting the spherical coordinates into second X-Y plane coordinates corresponding to the second image; determining whether the second X-Y plane coordinates of the end points all are located in the second image; and if the second X-Y plane coordinates of the end points all are located in the second image, generating a second rectangular mask in the second image by a maximum X coordinate value, a maximum Y coordinate value, a minimum X coordinate value and a minimum Y coordinate value of the second X-Y plane coordinates of the end points.