Robot Vision Measurement Parameter Tuning Without Pre-Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for optimizing measurement parameters in robot systems for object measurement are time-consuming and require significant effort, with a high dependency on mechanical characteristics of the robot, making adjustments difficult, especially when the evaluation target changes.
Innovation Solution
A method that involves acquiring multiple images of objects under varying conditions to estimate evaluation values for measurement parameters, allowing for the selection of optimized parameters based on predetermined criteria without the need for extensive pre-measurement, thus reducing effort and time, and eliminating robot dependency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods are used to optimize measurement parameters by evaluating combinations of parameters such as object imaging speed, captured image count, and imaging time interval, then measurement parameter optimization can be achieved, but the adjustment and optimization process takes a lot of time and effort
Solution Approach 1:
The patent creates a simulation environment that copies the physical measurement system, allowing parameter optimization to be performed in the virtual space rather than through actual physical measurements. The simulation model replicates the behavior of the robot and measurement device, enabling evaluation of parameter combinations without time-consuming physical trials.
Solution Approach 2:
The patent performs measurement parameter optimization in advance through simulation before actual measurement tasks are executed. By pre-evaluating parameter combinations in the simulation environment, the system prepares optimized parameters beforehand, eliminating the need for time-consuming adjustments during actual operation.
2Measurement precision
If measurement parameters are adjusted for different robots based on their mechanical characteristics, then accurate measurement can be achieved, but the adjustment process becomes extremely difficult and requires significant effort
Solution Approach 1:
The patent changes the approach from manually adjusting parameters for each robot to automatically determining optimal parameters through simulation. The system evaluates multiple parameter combinations and selects the best ones based on simulation results, eliminating the need for manual adjustment and making the process easier to operate.
Solution Approach 2:
The simulation system automatically performs the parameter optimization process without requiring manual intervention. The system self-evaluates parameter combinations and self-selects optimal parameters, reducing the operational burden on users.
3Adaptability or versatility
If the evaluation target in object recognition is changed, then the measurement can adapt to different objects, but more effort and time are required for parameter adjustment and optimization
Solution Approach 1:
The simulation environment serves as a universal platform that can handle multiple evaluation targets and object types. By using the same simulation framework for different objects, the system maintains adaptability while avoiding the need for separate optimization processes for each object type.
Solution Approach 2:
The system performs comprehensive parameter optimization in advance through simulation that covers multiple evaluation targets. This preliminary optimization prepares the system to handle different objects without requiring additional time-consuming adjustments when the evaluation target changes.
Data Source
AI summary
A measurement parameter for use when measuring an object with a measuring device provided on a robot may be adjusted and optimized significantly more easily than in conventional technology. A measurement parameter optimization method or operations performed by a processor may include: acquiring N captured images of objects with first measurement parameters; estimating recognized object counts Zi for the objects based on acquiring N/j captured images of the objects with second measurement parameters, and storing the recognized object counts Zi as first data; based on acquiring N/j/k captured images of the objects with third measurement parameters, estimating recognized object counts Zi for the objects based on the first data and storing the recognized object counts Zi as second data; and determining an optimized measurement parameter that satisfies a predetermined judgment criterion from among the second data.