NC Program Search via Image Similarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing NC program searching methods are inefficient due to the need for generating machining element data from three-dimensional shape data and require labor-intensive input of NC programs and machining results, making it difficult to find matching programs.
Innovation Solution
An NC program searching method that generates image data for machining targets and finish shapes, calculates similarity between these images, and displays candidate NC programs for efficient retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machining element data is generated from three-dimensional shape data using conventional methods, then the searching unit can store and search programming basic data, but the search process becomes complex and requires manual setting of search conditions
Solution Approach 1:
The patent creates image data copies of workpiece shapes from three-dimensional shape data and stores them in an image data storage unit. This copying approach simplifies the search process by enabling direct visual pattern matching between target workpiece images and stored images, eliminating the need for complex machining element data generation and manual search condition setting while maintaining accurate shape comparison
Solution Approach 2:
The patent replaces the mechanical data processing system (generating machining element data from three-dimensional shape data) with an image-based system. By converting shape data into image data and using image comparison algorithms, the system substitutes complex computational mechanics with visual pattern recognition, simplifying the searching unit's operation while preserving search accuracy
2Loss of information
If the operator inputs NC programs and machining results manually for search, then the system can store program information, but the search process becomes labor-intensive and time-consuming
Solution Approach 1:
The patent performs preliminary actions by automatically generating image data from three-dimensional shape data and pre-storing it in the image data storage unit associated with corresponding NC programs. This preparation work is done in advance, so when a search is needed, the system can immediately compare target workpiece images with pre-stored images and retrieve relevant NC programs without requiring manual input or time-consuming processing
Solution Approach 2:
The system performs self-service by automatically generating image data, storing it with associated NC programs, and conducting image comparisons without operator intervention. The searching unit autonomously retrieves NC programs by comparing target workpiece images with stored images and identifying matches, eliminating the need for operators to manually input program information or search through databases
3Ease of manufacture
If NC programs are created for all products according to the same procedure, then program creation is systematic, but efficiency in creating NC programs is insufficient
Solution Approach 1:
The patent enables efficient program creation by copying and reusing NC programs based on image similarity. When a new workpiece is to be machined, the system compares its image with stored workpiece images, identifies similar products, and retrieves their NC programs for reuse. This allows systematic program creation through copying and adaptation rather than creating programs from scratch for each product
Solution Approach 2:
The patent makes the NC program creation system universal by enabling a single NC program to be adapted for multiple similar products. The image-based search system identifies products with similar geometries and retrieves their NC programs, allowing the same program creation methodology to serve multiple product types efficiently rather than requiring separate procedures for each product
Data Source
AI summary
In the invention, there are included a first image generating that generates image data corresponding to a shape data of a machining target as machining target image data, a second image generating that generates, based on NC programs created in the past, image data of a machining finish shape as finish shape image data, a storing that stores the finish shape image data associating with an NC program corresponding thereto, a similarity calculating that compares the machining target image data and the finish shape image data to calculate similarity of both the image data, a searching that searches, based on the similarity, candidates of finish shape image data similar to or matching the machining target image data and displays the candidates, and a first program displaying that displays, out of the candidates of the finish shape image data, an NC program for designated finish shape image data.


