Object Detection Using One-Dimensional Profile Screening
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing obstacle detection systems face challenges in efficiently detecting objects, such as pedestrians, on roadways due to high computational requirements and complexity in image processing, especially when dealing with varying light conditions and complex backgrounds.
Innovation Solution
A method that generates a one-dimensional profile from a grayscale image by summing pixel values column-wise and applying a low-pass filter to identify vertical edges, allowing for direct object detection within a defined narrow horizontal strip below the horizon, thereby reducing computational effort and enhancing robustness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a full grayscale image is processed for object detection, then detection accuracy is maintained, but computational effort and processing time increase significantly
Solution Approach 1:
The patent divides the image processing task into two stages: first, a rapid one-dimensional profile analysis is performed on the entire image to identify regions of interest; second, full grayscale image processing is applied only to these specific regions. This segmentation allows the system to maintain high detection accuracy while significantly reducing overall computational effort by avoiding exhaustive processing of the entire image.
Solution Approach 2:
The patent applies partial action by using a simplified one-dimensional profile method as a preliminary screening step before applying the more computationally intensive full grayscale image processing. This partial processing approach identifies candidate regions that require detailed analysis, thereby reducing the total computational load while maintaining detection accuracy for objects of interest.
2Reliability
If the entire image area is searched for objects, then no objects are missed, but computational time and resources increase
Solution Approach 1:
The patent performs preliminary action by generating one-dimensional profiles and identifying regions of interest before conducting full object detection. This preliminary screening step quickly eliminates large portions of the image that do not contain objects, allowing the system to maintain high detection reliability while significantly reducing the time required to process the entire image area.
3Measurement precision
If complex image processing algorithms are used, then detection accuracy improves, but device complexity and computational requirements increase
Solution Approach 1:
The patent segments the detection process into a simple one-dimensional profile analysis stage and a more complex grayscale image processing stage applied only to regions of interest. This segmentation reduces overall device complexity by limiting the application of complex algorithms to small portions of the image, while still achieving high detection accuracy through the combination of both stages.
Solution Approach 2:
The patent extracts regions of interest from the full image based on one-dimensional profile analysis before applying complex detection algorithms. This extraction approach removes the need to apply complex processing to the entire image, thereby reducing device complexity and computational requirements while maintaining detection accuracy for the extracted regions.
Data Source
Figure 1
Figure 2~4
Figure 5
AI summary
A method is described for detecting an object, in particular a pedestrian, located on a roadway, in the vicinity within the field of view of an optical sensor mounted on a support such as a vehicle, in which a relevant spatial area below the horizon is determined from the field of view of the optical sensor, a grayscale image is generated by means of the optical sensor which comprises a relevant image area corresponding to the relevant spatial area, and for the purpose of detecting the object, only this relevant image area corresponding to the relevant spatial area below the horizon is searched for a possible object.