Medium Conveyance Apparatus End Part Detection via Image Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medium conveyance apparatuses face challenges in precisely detecting the end parts of a medium from an image, especially when the medium thickness is not constant, leading to inaccuracies in image processing and medium identification.
Innovation Solution
The apparatus includes a conveying module, an imaging device, an edge pixel detection module, and an end part detection module, which utilize a reference member and fluctuation region detection to accurately identify the end parts of the medium by distinguishing between reliable and unreliable regions within the image, thereby enhancing detection precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If the imaging device captures the entire medium including regions with varying thickness, then the coverage area is improved, but the detection precision of end parts deteriorates due to noise from unreliable regions
Solution Approach 1:
The patent divides the input image into multiple regions: reliable regions and unreliable regions. The unreliable regions are identified based on the positional relationship between the imaging device and the conveying module, where medium thickness variations cause detection errors. By segmenting the image and selectively processing only reliable regions for end part detection, the system maintains high detection precision while preserving comprehensive image coverage.
Solution Approach 2:
The patent applies different processing qualities to different parts of the image. Reliable regions undergo edge pixel detection and end part identification, while unreliable regions are excluded from this processing. This local quality approach ensures that detection operations are performed only where they will yield accurate results, preventing noise from varying medium thickness from degrading overall detection precision.
2Reliability
If edge pixel detection is performed across the entire input image, then the detection completeness is improved, but the detection accuracy deteriorates due to inclusion of low reliability regions
Solution Approach 1:
The patent performs preliminary identification of unreliable regions before conducting edge pixel detection. By pre-determining which regions of the input image are unreliable (based on the imaging device's position relative to the conveying module), the system can exclude these regions from subsequent edge detection operations. This preliminary action ensures that only reliable regions contribute to end part detection, maintaining both completeness and accuracy.
Solution Approach 2:
The patent extracts and excludes unreliable regions from the set of regions subjected to edge pixel detection. By taking out the low reliability regions from the processing pipeline, the system prevents these areas from introducing detection errors while maintaining detection completeness across all reliable regions.
Data Source
AI summary
Provided are a medium conveyance apparatus, control method, and control program is to enable more precisely detecting an end part in a main scan direction of a medium from an image. A medium conveyance apparatus includes a conveying module to convey a medium, an imaging device to capture an image of the conveyed medium, a storage device to store a low reliability region inside an input image of a medium captured by the imaging device based on a positional relationship between an imaging position of the imaging device and arrangement position of the conveying module, an edge pixel detection module to detect edge pixels from the input image, an end part detection module to detect an end part in a main scan direction of the medium based on edge pixels detected from a region not including the low reliability region inside the input image, and an output control module to output information relating to the detected end part.