Multi-Sensor Camera System for Object Flow Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera systems struggle to maintain high image quality when combining data from multiple sensors, particularly in logistics automation, where objects are conveyed past code readers, leading to overlapping fields of view and complex image stitching challenges.
Innovation Solution
A camera system and method that determine regions of interest (ROI) and use image data from a single source within these areas, allowing for seamless assembly of images while minimizing quality loss, by using line or matrix sensors and a geometry detection sensor to define areas of interest and combine data in a way that avoids impairments in critical image areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple sensors are arranged next to one another to cover a broader reading range, then the reading width increases, but image quality deteriorates in overlapping areas due to stitching artifacts
Solution Approach 1:
The patent divides the reading area into multiple detection areas, each covered by a separate detection unit (camera). Each detection unit independently captures images of objects within its detection area, avoiding the need to stitch images from overlapping regions. This segmentation approach maintains high image quality while achieving broad reading width coverage.
Solution Approach 2:
The patent introduces an evaluation unit that acts as an intermediary between multiple detection units and the output system. This evaluation unit receives image data from multiple detection units, determines regions of interest, and selects or combines image data to generate a final output image. The evaluation unit ensures that regions of interest are captured with high quality by selecting the best image data from appropriate detection units.
2Area of stationary object
If image data from multiple sources is combined through stitching, then the overall image coverage increases, but image quality in critical areas deteriorates due to stitching impairments
Solution Approach 1:
The patent applies different processing strategies to different regions of the image. For regions of interest (such as code areas or text fields), the system ensures high image quality by selecting image data from a single detection unit or by prioritizing quality over seamless stitching. For non-critical areas, standard stitching or combination methods can be applied. This local quality approach ensures that critical areas maintain high image quality while still achieving comprehensive coverage.
Solution Approach 2:
The evaluation unit determines regions of interest before final image assembly. By identifying areas containing codes, text, or other critical information in advance, the system can prioritize these regions during the image selection and combination process, ensuring they are captured with the highest possible quality from the most appropriate detection unit.
3Manufacturing precision
If cameras are aligned and calibrated very precisely to control the recording situation, then stitching quality improves, but system complexity and setup difficulty increase
Solution Approach 1:
The evaluation unit automatically determines regions of interest and selects appropriate image data from multiple detection units without requiring manual alignment or calibration of the cameras. The system self-adjusts by evaluating the content and quality of image data from each detection unit and automatically selecting the best data for regions of interest. This eliminates the need for complex manual setup and calibration procedures.
4Device complexity
If a single sensor is used to record all relevant information, then system complexity is reduced, but the field of view is insufficient to cover the entire width
Solution Approach 1:
The patent combines the capabilities of multiple detection units (cameras) to achieve a wide field of view that covers the entire conveyor belt width. Each detection unit covers a portion of the conveyor belt, and the evaluation unit merges the image data from these multiple sources. This merging approach provides comprehensive coverage while maintaining relatively simple individual detection units.
Data Source
Figure 1~2
Figure 3
AI summary
A camera system (10) for capturing a stream of objects (14) moving relative to the camera system (10) is described. The camera system (10) comprises several capture units (18a-b), each having an image sensor for acquiring image data from a capture area (20a-b) that partially overlaps and together covers the width of the stream of objects (14), and an evaluation unit (28) for assembling image data from the capture units (18a-b) into a common image and for identifying areas of interest within the image data. The evaluation unit (28) is configured to use only image data from the same capture unit (18a-b) for the common image within an area of interest when assembling it.