Hough Space Time Sequence Pattern Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for real-time similarity matching of industrial time sequence data are not effective for visualizing and matching one-dimensional or multi-dimensional data, as they lack specificity and efficiency in industrial applications.
Innovation Solution
A visualized time sequence pattern matching method based on Hough transformation, which involves judging data type, normalization, time sequence selection, Hough transformation, and a voting mechanism to convert original coordinates into Hough space for similarity matching, allowing for efficient identification of similar time sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional pattern matching methods are used for industrial time sequence data, then the matching process can be performed, but the efficiency and effectiveness for real-time visualization matching is insufficient
Solution Approach 1:
The patent transforms one-dimensional time sequence data into two-dimensional coordinate space by treating time as the X-axis and data values as the Y-axis. This dimensional transformation enables the application of Hough transformation algorithms originally designed for 2D image processing, thereby improving both matching efficiency and effectiveness for time sequence data visualization.
Solution Approach 2:
The patent replaces traditional mechanical pattern matching approaches with Hough transformation, a mathematical transformation method. By substituting the matching mechanism with a transformation-based approach in Hough space, the system achieves faster and more reliable pattern recognition in time sequence data.
2Adaptability or versatility
If Hough transformation is applied to time sequence data, then visualization matching capability is improved, but the method lacks specificity for industrial time sequence data characteristics
Solution Approach 1:
The patent introduces local quality control through the parameter λ (lambda), which dynamically adjusts the width of the voting interval in Hough space based on local data characteristics. This allows the method to adapt to different segments of time sequence data with varying properties, maintaining both versatility and specificity for industrial applications.
Solution Approach 2:
The patent modifies the standard Hough transformation by introducing adjustable parameters including λ (voting interval width), threshold values for similarity judgment, and normalization parameters. These parameter changes enable the method to be specifically tuned for industrial time sequence data while maintaining the core Hough transformation framework.
3Adaptability or versatility
If multi-dimensional time sequence data is processed, then comprehensive analysis is achieved, but normalization processing is required which increases computational complexity
Solution Approach 1:
The patent implements a universal normalization framework that handles both one-dimensional and multi-dimensional time sequence data through the same Hough transformation process. The normalization parameters and voting mechanism work consistently across different data dimensions, reducing the need for dimension-specific processing logic despite the increased computational requirements.
Data Source
AI summary
The invention provides a visualized time sequence pattern matching method based on Hough transformation, and relates to the technical field of data visualization analysis. The method comprises the steps of: firstly, judging whether historical data to be matched is one-dimensional time sequence data or multi-dimensional time sequence data, and if the historical data to be matched is the multi-dimensional time sequence data, performing normalization processing; performing time sequence selection: selecting a time sequence to be matched from the historical data in a time window pattern, and eliminating the selected time sequence from the historical data; converting a time sequence image in original coordinates to Hough space through the Hough transformation, and judging the similarity matching situation of the time sequence through a voting mechanism; and finally, screening the finally-matched results according to the voting results.