A solder joint inspection model training method includes the steps of: training a first identification model according to first sample images to identify a surface-mount device with a solder joint in an image; training a second identification model according to second sample images to identify a surface-mount device without a solder joint in an image; inputting labeled original images to a trained first identification model to output first images; inputting the first images to a trained second identification model to output second images; masking the first images with the second images to generate images with normal solder joints and images with abnormal solder joints; and training a solder joint inspection model based on the images with normal solder joints and the images with abnormal solder joints.

Systems and methods for real time feedback and for updating welding instructions for a welding robot in real time is described herein. The data of a workspace that includes a part to be welded can be received via at least one sensor. This data can be transformed into a point cloud data representing a three-dimensional surface of the part. A desired state indicative of a desired position of at least a portion of the welding robot with respect to the part can be identified. An estimated state indicative of an estimated position of at least the portion of the welding robot with respect to the part can be compared to the desired state. The welding instructions can be updated based on the comparison.

A computer-implemented method for operating a laser material processing machine. An estimated result is ascertained as a function of predefined process parameters, which characterize how good an actual result of the laser material processing will be, and the process parameters are varied by means of Bayesian optimization with the aid of a data-based model, until an actual result of the laser material processing is sufficient enough.

A system for monitoring a laser machining process for machining a workpiece includes a computing unit configured to determine an input tensor on the basis of current data of the laser machining process and to determine an output tensor on the basis of the input tensor using a transfer function. The output tensor contains information on a current machining result. The transfer function between the input tensor and the output tensor is formed by a trained neural network.

A system for detecting machining errors for a laser machining system for machining a workpiece includes: a detection unit for detecting image data and height data of a machined workpiece surface; and a computing unit. The computing unit is designed to generate an input tensor based on the detected image data and height data and to determine an output tensor on the basis of the input tensor using a transfer function. The output tensor contains information on a machining error.

A computer-implemented method determining one or more parameter values for a laser-engraving process, the method comprising: executing a laser pulse model on a computer-simulated surface to generate a first surface geometry on the computer-simulated surface, wherein the laser pulse model is based on a first set of values for a set of parameters; determining a quality score for the first surface geometry; based on the quality score, performing a global optimization process to generate a second set of values for the set of parameters; and modifying the laser pulse model based on the second set of values to generate a modified laser pulse model.

A laser machining apparatus includes an actuator that changes relative positions of a machining head and a workpiece; a control unit that controls in machining execution the laser oscillator, the machining head, and the actuator based on a machining parameter; a machining state observation unit that detects, from process light that is light generated from the workpiece by laser beam irradiation, light intensities in a plurality of predetermined wavelength bands as a plurality of optical sensor signals; a feature extraction unit that extracts at least one of features, the features being obtainable from an index of correlation between the plurality of optical sensor signals and from one of the optical sensor signals; and a correction quantity calculation unit that determines the machining parameter to be corrected as a correction parameter and a correction quantity for the correction parameter based on the at least one of the features.

A system for detecting machining errors for a laser machining system for machining a workpiece includes: a detection unit for detecting image data and height data of a machined workpiece surface; and a computing unit. The computing unit is designed to generate an input tensor based on the detected image data and height data and to determine an output tensor on the basis of the input tensor using a transfer function. The output tensor contains information on a machining error.

A method of inspecting an electric motor includes scanning an electric motor stator winding with a 2D or 3D camera, acquiring one or more images of a plurality welds between adjacent electrical wires forming the stator winding using the 2D camera, analyzing the one or more acquired images with at least one neural network such that the neural network determines if at least one of the plurality of welds has a weld defect. The at least one neural network is trained and distinguishes between surface discoloration on a surface of the welds and defect discoloration resulting from contamination during welding. Also, the method inspects over 150 welds per electric motor stator winding moving along an assembly line.

Systems and methods for real time feedback and for updating welding instructions for a welding robot in real time is described herein. The data of a workspace that includes a part to be welded can be received via at least one sensor. This data can be transformed into a point cloud data representing a three-dimensional surface of the part. A desired state indicative of a desired position of at least a portion of the welding robot with respect to the part can be identified. An estimated state indicative of an estimated position of at least the portion of the welding robot with respect to the part can be compared to the desired state. The welding instructions can be updated based on the comparison.