Dual-Camera Zoom Object Detection With Cross-View Image Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic devices struggle to accurately analyze images and identify objects when there is insufficient information, such as when images are taken with a zoomed-in camera, missing parts of the object, leading to incomplete object recognition.
Innovation Solution
The electronic device uses multiple cameras, one in zoom mode and one in normal mode, to obtain images, detect objects, and perform object recognition by correlating the images to accurately identify objects even when zoomed-in images are incomplete.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Length of moving object
If a camera operates in zoom mode to capture distant objects, then the ability to capture distant objects is improved, but the completeness of object information is worsened because parts of the object may be missed
Solution Approach 1:
The patent divides the imaging task into two segments: one camera captures the overall scene context in normal mode, while another camera captures detailed zoomed-in views. This segmentation allows each camera to specialize in different aspects (context vs. detail), resolving the contradiction between zoom capability and information completeness.
Solution Approach 2:
The processor acts as an intermediary that receives images from both cameras, identifies corresponding regions between the normal mode and zoom mode images, and synthesizes complete object information by combining data from both sources, thereby compensating for information loss in zoomed-in views.
2Measurement precision
If multiple cameras are used to capture complete object information, then object recognition accuracy is improved, but device complexity is worsened
Solution Approach 1:
Both cameras in the system perform multiple functions: they capture images for object recognition, provide contextual information, and enable each other's data to be processed through the processor's image identification algorithm. This multi-functionality justifies the added complexity by maximizing the utility of each camera component.
3Manufacturing precision
If zoom mode is used to capture detailed views, then detail visibility is improved, but the ability to recognize complete objects is worsened when parts are missing
Solution Approach 1:
The system performs preliminary actions by capturing normal mode images before zooming in, establishing a complete reference view of the scene. This preliminary comprehensive view allows the processor to later identify and fill in missing object parts when analyzing zoomed-in details, ensuring reliable object recognition even when zoomed views are incomplete.
Data Source
AI summary
An electronic device and a controlling method thereof are provided. The electronic device includes a display, a first camera, a second camera, memory storing one or more computer programs, and one or more processors communicatively coupled to the display, the first camera, the second camera, and the memory, wherein the one or more computer programs include computer-executable instructions that, when executed by the one or more processors, cause the electronic device to, obtain a first image through the first camera operating in a zoom mode, control the display to display the first image, obtain a second image through the second camera operating in a normal mode, identify a third image of the first camera in the normal mode corresponding to the first image based on a zoom-in ratio in the zoom mode, by detecting an object included in the second image, obtain location information regarding the detected object in the second image, based on a relative location between the second image and the third image, obtain location information of the object in the third image corresponding to the location information obtained from the second image, and, based on the location information of the object, detect the object in the first image.