Object Detection via Pixel Column Depth Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques for object detection using depth information are inefficient, unreliable due to image imperfections, and lack accuracy.
Innovation Solution
A method that fits equations to pixel columns in an image to detect objects, splits objects if the fit is insufficient, and merges objects with similar depth and equation profiles, using a stereoscopic camera system to generate depth maps and object probability maps for accurate object detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional object detection techniques are used, then object detection can be performed, but computational efficiency is poor and accuracy is insufficient
Solution Approach 1:
The image is divided into multiple pixel columns for independent processing. Each pixel column is analyzed separately to determine object presence and depth, allowing parallel computation and improving both efficiency and accuracy. The segmentation of the detection space into discrete columns enables systematic processing of depth information.
Solution Approach 2:
The patent introduces depth as an additional dimension beyond traditional 2D image analysis. By incorporating depth information from pixel columns, the system creates a 3D spatial understanding of objects, improving detection accuracy through volumetric analysis rather than planar analysis alone.
2Reliability
If traditional object detection techniques are used, then detection can be performed, but reliability is poor when images have imperfections such as noisy or missing pixels
Solution Approach 1:
The system employs iterative refinement where detected objects are used to generate expected depth profiles, which are then compared with actual pixel column depths. Discrepancies feed back into the detection process, allowing the system to correct errors and improve reliability even when initial detections are affected by noisy or missing pixels.
Solution Approach 2:
The patent changes the parameter space by using depth information from pixel columns as an additional verification parameter. Instead of relying solely on 2D image features, the system uses depth consistency across multiple pixel columns to confirm object presence and location, making detection more robust to image imperfections.
3Reliability
If objects are detected using depth information, then detection capability is improved, but computational complexity increases
Solution Approach 1:
By segmenting the image into pixel columns and processing each independently, the system reduces the complexity of analyzing the entire image at once. Each pixel column can be processed in parallel, and the results are combined to form the final detection, simplifying the overall computational structure while maintaining reliability.
Solution Approach 2:
The patent extracts depth information specifically from pixel columns rather than processing the entire image data. This extraction of relevant depth parameters from selected pixel columns reduces computational complexity while preserving the reliability benefits of depth-based detection.
Data Source
AI summary
First and second objects are detected within an image. The first object includes first pixel columns, and the second object includes second pixel columns. A rightmost one of the first pixel columns is adjacent to a leftmost one of the second pixel columns. A first equation is fitted to respective depths of the first pixel columns, and a first depth is computed of the rightmost one of the first pixel columns in response to the first equation. A second equation is fitted to respective depths of the second pixel columns, and a second depth is computed of the leftmost one of the second pixel columns in response to the second equation. The first and second objects are merged in response to the first and second depths being sufficiently similar to one another, and in response to the first and second equations being sufficiently similar to one another.