Camera Apparatus for Object Image Stitching Using Geometrical Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera systems struggle to efficiently generate high-quality object images from a moving stream of objects on a conveyor belt, particularly when multiple sensors are arranged next to each other, as they face challenges in stitching images due to high data volume and complexity, leading to difficulties in obtaining individual object images for processing.
Innovation Solution
A camera apparatus and method that utilize overlapping frames from multiple cameras, supported by geometrical data from sensors like laser scanners or 3D cameras, to stitch object images directly from frames, simplifying the image processing and ensuring only relevant frames are used, thereby reducing data volume and processing complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple sensors are arranged next to each other to cover a wider reading region, then the coverage area is improved, but the data volume and processing complexity increase significantly
Solution Approach 1:
The patent divides the wide reading region into multiple sub-regions, each covered by a separate sensor with a narrow field of view. Each sensor processes only its local region independently, avoiding the need to process the entire wide area simultaneously. This segmentation reduces the data volume each sensor must handle and simplifies the overall processing architecture.
Solution Approach 2:
The patent introduces an intermediary processing step that selectively extracts and combines relevant image data from multiple sensors. Instead of directly processing all sensor outputs, the system uses intermediate processing to identify and stitch together only the necessary regions, reducing complexity while maintaining comprehensive coverage.
2Area of stationary object
If multiple sensors are arranged next to each other to cover a wider reading region, then the coverage area is improved, but the data volume increases to a huge amount
Solution Approach 1:
The patent extracts only the relevant portions of image data from each sensor's field of view, rather than processing complete images from all sensors. By taking out and processing only the necessary regions of interest, the system significantly reduces the total data volume while maintaining comprehensive coverage of the wide reading region.
Solution Approach 2:
The patent implements partial processing by focusing computational resources on specific regions of interest within each sensor's coverage area. Instead of processing every pixel from all sensors equally, the system applies partial action to only those areas containing relevant objects or information, reducing overall data volume and processing requirements.
3Area of stationary object
If image stitching methods are used to join images from multiple sensors, then the coverage is improved, but the quality and success of stitching greatly depend on image data and become complex with more sensors
Solution Approach 1:
The patent performs preliminary actions by pre-aligning and pre-processing images from multiple sensors before the actual stitching operation. Calibration and coordinate transformation are applied in advance to ensure that images from different sensors are properly oriented and scaled, making the subsequent stitching process more reliable and less dependent on the quality variations of individual images.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for robust, efficient, and high-quality object image generation, enabling real-time processing and maintaining legibility of sensitive regions like code areas, while simplifying image fusion and segmentation, and facilitating the creation of 3D models with textures.
Implementation Method 1
a laser scanner or a 3D camera is arranged upstream of the individual cameras in the direction of movement of the stream of objects with respect to the camera apparatus to generate geometrical data of the objects
Data Source
AI summary
A camera apparatus is provided for detecting a stream of objects moving relative to the camera apparatus having a plurality of individual cameras that each have an image sensor for recording frames, wherein the frames overlap one another in part, having an evaluation unit for compiling frames, and having a geometry detection sensor for detecting geometrical data of the objects. The evaluation unit is here configured to generate an object image assembled from frames of an individual object of the stream of objects, with the selection of the participating frames and/or the assembly taking place on the basis of the geometrical data.