Hybrid MB/ML techniques for automated PCB defect detection comprising an inspection module including an autoencoder trained to recognize at least one feature of interest, and a defect detection module

A computer program product including one or more non-transitory machine-readable mediums having instructions encoded thereon that when executed by at least one processor cause a process to be carried out for detecting defects in one or more features of interest on a printed circuit board (PCB), the process comprising:

• processing an image of the PCB using an autoencoder to generate a reconstructed image, the autoencoder having been trained to recognize the at least one feature of interest based on a combination of model data and training data, wherein images corresponding to the model data are compressed or encoded, using an encoder of the autoencoder, into a latent-space representation, and images corresponding to the training data are compressed and encoded into the latent-space representation using the encoder, and wherein, in the latent-space representation, shared representations of encoded model-based features and machine learning features are analyzed to identify the correlations between the features from each processing chain, wherein the model data comprises various pictorial representations of PCB components and layouts, including drawings, sketches, models, and/or other characterizations or representations, including a computer-assisted drawing (CAD) image, rather than photographic images of PCBs or their components;
• comparing the reconstructed image to the image;
• based on the comparison, generating a reconstruction error value representing a measure of difference between the image and the reconstructed image;
• and
• generating an output indicating detection of a defect in at least one of the one or more features of interest based on the reconstruction error value transgressing a threshold.

A method of detecting defects in one or more features of interest on a printed circuit board (PCB), the method comprising:

• processing an image of the PCB using an autoencoder to generate a reconstructed image, the autoencoder having been trained to recognize the at least one feature of interest based on a combination of model data and training data, wherein images corresponding to the model data are compressed or encoded, using an encoder of the autoencoder, into a latent-space representation, and images corresponding to the training data are compressed and encoded into the latent-space representation using the encoder, and wherein, in the latent-space representation, shared representations of encoded model-based features and machine learning features are learned to discover the correlations between the features from each processing chain, wherein the model data comprises various pictorial representations of PCB components and layouts, including drawings, sketches, models, and/or other characterizations or representations, including a computer-assisted drawing (CAD) image, rather than photographic images of PCBs or their components;
• comparing the reconstructed image to the image;
• based on the comparison, generating a reconstruction error value representing a measure of difference between the image and the reconstructed image;
• and
• generating an output indicating detection of a defect in at least one of the one or more features of interest based on the reconstruction error value transgressing a threshold.