Systems and methods for recognition of unreadable characters on printed circuit boards
AssigneeSkyworks Solutions, Inc.
InventorCarlos Fabian Nava
Life of this patent
Estimating the term…
Abstract
Systems and methods for recognition of unreadable characters on printed circuit boards. In some embodiments, a method for recognizing characters can be utilized for recognition of damaged characters on a printed circuit board. The method can include obtaining a digital image for each of a plurality of characters on the printed circuit board. The method can further include dividing each digital image into an array of regions. The method can further include generating a data structure from the arrays of the digital images. The data structure can include gradient features based on stroke shapes on small distances, structural features based on stroke trajectories on extended distances, and concavity features based on stroke relationships.
Claims16
A method for recognizing characters imprinted on a printed circuit board, the method comprising:
- obtaining a digital image for each of a plurality of characters on the printed circuit board;
- dividing each digital image into an array of regions;
- and
- generating a data structure from the arrays of the digital images, the data structure including gradient features based on stroke shapes, structural features based on stroke trajectories, and concavity features based on stroke relationships, the gradient features, the structural features, and the concavity features being obtained with a gradient, structural and concavity (GSC) algorithm that includes or supports one or more of (i) two dimensional convolutions of each digital image with Sobel operators to obtain the gradient features, (ii) addition of zero padded borders to the array of each digital image and seeking gradient patterns among array elements immediately surrounding each non-padded border array element to obtain the structural features, and (iii) generation of a coarse pixel density of the array of each digital image with the coarse pixel density including a threshold applied to a count of all character pixels among the respective array.
A method for recognizing characters imprinted on a printed circuit board, the method comprising:
- obtaining a digital image of a character on the printed circuit board;
- and
- performing an optical character recognition process on the digital image using trained information that includes gradient features based on stroke shapes, structural features based on stroke trajectories, and concavity features based on stroke relationships, the gradient features, the structural features, and the concavity features being obtained with a gradient, structural and concavity (GSC) algorithm that includes two dimensional convolutions of the digital image with Sobel operators to obtain the gradient features.
A system for processing printed circuit boards, comprising:
- a first production system configured to fabricate a plurality of printed circuit boards, each printed circuit board including a plurality of characters imprinted thereon to provide information about the printed circuit board;
- and
- a character recognition system configured to monitor at least some of the printed circuit boards, and including an imaging component configured to obtain a digital image of the characters on a selected printed circuit board, and a processor configured to perform an optical character recognition process on the digital image using trained information that includes gradient features based on stroke shapes, structural features based on stroke trajectories, and concavity features based on stroke relationships, the gradient features, the structural features, and the concavity features being obtained with a gradient, structural and concavity (GSC) algorithm that includes two dimensional convolutions of the digital image with Sobel operators to obtain the gradient features.
Dependencies are inferred from the claim text.