Image Pattern Recognition for Mail Sorting Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mail sorting systems face challenges in efficiently identifying and sorting mail items when the recipient address is invalid or difficult to read, leading to issues in delivering mail items to the correct sender for return.
Innovation Solution
A method and device that utilize an image pattern recognition system, including a classification device, sorting system, and image pattern data storage, which undergo a training phase to identify and store image patterns, allowing for automatic sorting without manual input or processor intervention, enabling efficient recognition and grouping of mail items by sender image patterns during transport.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual input and processor intervention are used to identify and sort mail items, then sorting accuracy can be maintained, but productivity decreases and time consumption increases
Solution Approach 1:
The system enables self-service automation by training the image pattern recognition system with sample mail items to automatically learn and identify sender image patterns. The system sorts mail items autonomously by comparing captured images against stored patterns, eliminating the need for manual input or processor intervention while maintaining high sorting accuracy and productivity
Solution Approach 2:
The system performs preliminary training action by capturing and storing image patterns from sample mail items before actual sorting operations. This pre-learning phase enables the system to quickly and accurately identify sender patterns during sorting without requiring manual intervention, thus improving productivity while maintaining automation
2Reliability
If traditional sorting systems re-sort mail items twice to handle undeliverable items, then return routing can be achieved, but time loss and productivity decrease
Solution Approach 1:
The system performs preliminary capture and storage of sender image patterns during the first sorting pass. When mail items become undeliverable, the pre-stored patterns enable immediate identification and routing during the return process, eliminating the need for a complete second sorting cycle and reducing time loss while maintaining reliable return routing
Solution Approach 2:
The system replaces the mechanical re-sorting process with an optical recognition system that identifies sender patterns through image analysis. This substitution allows the system to quickly determine return destinations without physically re-sorting mail items through the entire sorting mechanism again, significantly reducing time loss while maintaining accurate return routing
3Ease of operation
If sender addresses are difficult to read from mail items, then automatic recognition fails, but manual reading increases operation complexity and time
Solution Approach 1:
The system replaces manual address reading with an optical image recognition system that captures and analyzes sender image patterns. The system can identify senders through their unique image patterns (such as logos, branding, or visual characteristics) even when text addresses are illegible or difficult to read, thus improving ease of operation without losing sender information
Solution Approach 2:
The system creates visual copies (images) of mail items and analyzes these copies to identify sender patterns. By working with image copies rather than directly reading physical addresses, the system can extract sender information from visual characteristics that may be more reliable than difficult-to-read text, maintaining information accuracy while improving operational ease
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a method and a device for transporting objects, in particular mail items, to destination points dependent on image patterns. In a training phase, a sample of objects to be transported passes through a classification device. An image acquisition device (1) generates at least one image of each object in the sample. A grouper determines, by applying a clustering method to the images, which different image patterns are represented on at least one object in the sample. These image patterns are stored in an image pattern data memory (4). In a sorting phase, each object to be transported (Ps-1, Ps-2, ...) passes through a sorting system. An image acquisition device (1) generates an image (Fig-1, Fig-2, ...) of the object (Ps-1, Ps-2, ...). The image pattern recognizer (5) evaluates the image (Fig-1, Fig-2, ...)The system automatically determines which stored image pattern is displayed on the object. Using this image pattern, the sorting system identifies a target point to which the object is to be transported and triggers the transport of the object (Ps-1, Ps-2, ...) to this target point.