Depth Map Based Partial Image Blurring for UAV Cameras
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques are unsuitable for cameras mounted on moving objects, such as unmanned aerial vehicles (UAVs), as they fail to effectively generate partially blurred images with a natural depth of field, especially for non-stereoscopic cameras.
Innovation Solution
The method involves generating a depth map of an image using pixel values representative of object distances, identifying different depths, and blurring pixels based on relative distances from a target depth, allowing for the creation of a partially blurred image even with 2D cameras on movable objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual pixel-blurring algorithm is applied to a single image, then processing can be done with simple camera systems, but the blurring appears unnatural and is not aesthetically appealing
Solution Approach 1:
The image is segmented into multiple depth layers using depth map generation. Each layer corresponds to objects at different distances from the camera, allowing selective blurring application based on depth information rather than uniform manual blurring across the entire image.
Solution Approach 2:
The patent introduces depth as an additional dimension by generating a depth map that provides three-dimensional distance information for each pixel. This enables the system to apply blurring based on spatial depth relationships, transforming the two-dimensional image processing into a three-dimensional depth-aware process that produces natural-looking depth of field effects.
2Difficulty of detecting and measuring
If multiple images from different perspectives are processed, then natural depth of field blur is achieved, but the technique is unsuitable for cameras that move significantly between captured images
Solution Approach 1:
The patent implements a dynamic approach by tracking features across multiple frames and continuously updating the depth map as the camera moves. The system adapts to camera motion by re-identifying features in new positions and recalculating depth information, allowing the partial blurring effect to be maintained throughout video sequences or continuous imaging scenarios.
Solution Approach 2:
The system uses feedback from feature tracking and depth map generation to continuously refine the blurring application. By monitoring feature positions across frames and updating depth information dynamically, the system adjusts the blurring parameters based on current scene geometry and camera position, ensuring natural-looking results even during significant camera movement.
3Difficulty of detecting and measuring
If depth map generation and depth-based blurring is implemented, then natural depth of field is achieved, but processing complexity increases
Solution Approach 1:
The patent introduces a depth map as an intermediary data structure that stores depth information for each pixel. This intermediate representation serves as a bridge between the input images and the final blurred output, enabling the system to apply complex depth-based blurring operations efficiently by referencing pre-computed depth values rather than performing complex geometric calculations during the actual blurring process.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
Methods and systems for processing an image having a first set of pixels. The method may include generating a depth map of the image, the depth map including a second set of pixel values representative of distances of objects in the image. The method may further include identifying a plurality of different depths at which objects are located in the image based on the depth map, and using the depth map to determine a relative distance between one identified depth in the plurality of different depths and each of the other identified depths in the plurality of different depths. Finally, the method may include blurring pixels in the first set of pixels based on each determined relative distance.