Component Mounter Search Range Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In the PP cycle of a component mounter, there is a challenge in improving the efficiency and precision of image processing for recognizing the supply state of components from a bulk feeder, especially when multiple components are picked up by multiple suction nozzles.
Innovation Solution
The component mounter is configured to target a predetermined search range based on the position and feature amount of a reference feature, improving the efficiency of search processing and preventing erroneous recognition of components by determining the candidate feature portions' belonging to the same component as the reference feature.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing is performed on the entire supply area to recognize component positions, then recognition precision is improved, but processing time increases and efficiency decreases
Solution Approach 1:
The patent divides the supply area into multiple search ranges, each centered on a reference feature portion. Instead of processing the entire supply area at once, the image processing is segmented into smaller regions, reducing the total processing time while maintaining recognition precision within each segment.
Solution Approach 2:
The patent performs preliminary identification of reference feature portions before conducting detailed component recognition. By first locating reference features and then using them to define search ranges, the system prepares the processing framework in advance, avoiding redundant calculations and improving overall processing efficiency.
2Reliability
If the search range is expanded to cover the entire supply area, then component detection completeness is improved, but processing complexity and time increase
Solution Approach 1:
The patent applies different processing strategies to different regions of the supply area. Reference feature portions receive special attention as they serve as anchors, while other areas are processed within constrained search ranges. This localized approach ensures detection completeness without uniformly applying complex processing to the entire area.
Solution Approach 2:
Reference feature portions act as intermediaries between the camera image and the component positions. These reference features facilitate the transformation from image coordinates to component locations, simplifying the overall detection process while maintaining completeness.
3Productivity
If feature portions from different components are mistakenly identified as belonging to the same component, then recognition speed is improved, but recognition precision deteriorates
Solution Approach 1:
The patent uses reference feature portions as feedback anchors to verify component boundaries. When identifying feature portions, the system continuously references the known positions of reference features to confirm whether detected features belong to the same component or a different one, preventing erroneous merges while maintaining efficient processing.
4Productivity
If multiple suction nozzles pick up multiple components simultaneously, then mounting productivity is improved, but the complexity of tracking and managing component positions increases
Solution Approach 1:
The patent segments the multi-component pickup process into individual component tracking within separate search ranges. Each component's feature portions are independently identified and tracked relative to its own search range defined by reference features, simplifying the management of multiple simultaneous pickups while maintaining high productivity.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
A component mounter includes a state recognition section configured to recognize a supply state of a component in a supply area based on a feature amount including at least one of a shape and an angle of each of multiple feature portions in image data on which image processing has been executed. The state recognition section determines whether one or more candidate feature portions included in a search range belong to an identical component together with a reference feature portion based on one of feature amounts of the candidate feature portion and a positional relationship between the reference feature portion and the candidate feature portion.