Fisheye Camera Image Analysis via Segmentation and Dewarping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fisheye camera images are severely distorted, making it difficult to accurately recognize objects using general intelligent image analysis methods, and existing solutions may fail to recognize objects on the boundaries of segmented images.
Innovation Solution
An image analysis apparatus that includes a pre-processing unit to segment and dewarp fisheye camera images, a post-processing unit to remove duplicate object detections, and a user interface to set field of view and image combination methods, allowing for accurate object recognition without significantly altering general image analyzers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fisheye camera images are segmented for analysis, then object recognition accuracy in distorted regions is improved, but objects located on boundaries between segmented images are not recognized or are incorrectly recognized
Solution Approach 1:
The fisheye camera image is divided into multiple segmented images with overlapping regions. Each segmented image is processed independently through dewarping and object detection, allowing accurate recognition of objects in distorted regions while the overlap ensures boundary objects are captured in multiple segments for verification
Solution Approach 2:
The segmented images are superimposed with overlapping regions before object detection. This preliminary action ensures that objects near boundaries are included in multiple segmented images, allowing the system to identify and remove duplicate detections in the post-processing stage
2Device complexity
If general intelligent image analysis methods are applied to fisheye camera images, then the analysis process is simple, but object recognition accuracy is poor due to severe distortion
Solution Approach 1:
The image is segmented into multiple regions, each with less severe distortion, allowing standard image analysis methods to work effectively on each segment while maintaining overall system simplicity
Solution Approach 2:
A preprocessing unit performs dewarping on each segmented image before object detection. This intermediary step corrects the severe distortion in each segment, enabling accurate object recognition without requiring complex distortion-invariant algorithms throughout the entire system
3Measurement precision
If segmented images are combined after dewarping, then object recognition accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
By dividing the image into segments, each with a smaller field of view, the dewarping computation for each segment is less intensive than processing the entire fisheye image at once, distributing the computational load
Solution Approach 2:
The segmented images are superimposed with overlapping regions and then combined using a preset combination method. This merging process integrates the results from multiple segments while the overlap ensures consistent object detection across boundaries
Data Source
AI summary
Disclosed is an image analysis apparatus for analyzing a camera image. The image analysis apparatus for analyzing a camera image segments an input fisheye camera image into segmented images with a preset size field of view and superimposes the segmented images so that some regions overlap, performs dewarping on each of the segmented images, then combines the segmented images on which the dewarping is performed using a preset combination method, generates an analysis image, and detects objects included in the analysis image. In this case, the image analysis apparatus removes a result recognized as a duplicate from a detection result of the object by post-processing.