Object Position Recognition Using Edge Deviation Minimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object positioning systems struggle to accurately recognize and position small objects with non-round shapes and those having chipped or inclined surfaces, as they require specialized illumination setups and are limited to specific shapes, leading to deviations and configuration challenges.
Innovation Solution
An object positioning system that captures images of objects from a direction perpendicular to their flat face, using an illuminating unit parallel to the optical axis, and performs image processing to set recognition lines, detect brightness transitions, and determine object outlines by minimizing edge point deviations, allowing for precise positioning of objects with flat faces, even those with chipped portions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If image processing is used to recognize object position based on outer shape, then recognition capability is provided, but measurement precision deteriorates due to hole wall portions in captured images causing deviation
Solution Approach 1:
The patent introduces a coordinate conversion intermediary that transforms captured image coordinates into object coordinate system. This conversion process acts as a mediator between the imperfect captured image (with hole wall deviations) and the accurate object position recognition, eliminating the measurement errors caused by optical system limitations.
2Adaptability or versatility
If special illumination devices are used to illuminate concave/convex portions, then recognition capability is improved, but device complexity increases and configuration becomes difficult for steep slopes
Solution Approach 1:
The patent employs self-illumination where the object's own reflective properties serve the illumination function. By using the object's inherent reflectivity characteristics and the optical system's capabilities, the solution eliminates the need for complex external illumination devices while still enabling recognition of concave and convex portions.
3Measurement precision
If recognition is limited to round holes with specific methods, then measurement precision is maintained, but adaptability deteriorates as line position detection in outer shape becomes impossible
Solution Approach 1:
The patent creates a universal recognition system that can handle multiple object types and shapes through coordinate conversion. The same processing framework accurately detects both round hole centers and line positions in outer shapes, eliminating the need for shape-specific methods and enabling versatile application across different object geometries.
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
Enables precise recognition and positioning of small objects with flat faces, including those with chipped or inclined surfaces, without the need for specialized illumination, improving accuracy and adaptability across various object shapes and sizes.
Implementation Method 1
an illuminating unit that illuminates the flat face of the object from a direction parallel to an optical axis of the image capturing unit
Data Source
AI summary
A system recognizes the outline of an object that includes, at its edge, a portion including a rollover or a chipped portion. An image processing unit finds an outline of an object having a flat face from an image captured perpendicular to the flat face. A dark-transition-boundary detecting unit detects, on each of a plurality of recognition lines, a possible boundary point of dark-transition where a bright-to-dark transition occurs from outside toward inside of the object. A bright-transition-boundary detecting unit detects a possible boundary point of bright-transition where a dark-to-bright transition occurs from outside toward inside of the object. An edge detecting unit detects an edge point on the basis of the possible dark-transition-boundary point and the possible bright-transition-boundary point. An outline-determining unit determines an outline of the object that minimizes the sum of deviations between the outline and the respective edge points detected on the recognition lines.