Sewing machine and methods of using the same

A sewing machine, comprising:

• a feed mechanism arranged in a sewing bed of the sewing machine wherein a workpiece can be placed on and be at least one of moved and rotated on the sewing bed by the feed mechanism, wherein the workpiece comprises two pieces of fabric joined together;
• a sewing head arranged above the sewing bed, the sewing head comprising:
• a needle bar extending toward the sewing bed to a distal end to which a needle is attached, wherein a thread is threaded through the needle;
• a motor for moving the needle bar in a reciprocating motion to move the needle and the thread through the workpiece to form one or more stitches in the workpiece during a sewing operation;
• and
• a swing motor for laterally swinging the needle bar during a sewing operation;
• a camera having a field of view encompassing at least a portion of the workpiece and any of the thread and the one or more stitches formed in the workpiece, wherein the camera generates a camera data signal related to the portions of the workpiece, the thread, and the one or more stitches formed in the workpiece in the field of view of the camera;
• an object recognition neural network that is trained to detect and to classify one or more features of the workpiece from the camera data signal as a recognized feature, wherein the object recognition neural network generates a feature detection data signal related to at least one of a location and a shape of a seam, an edge, a pocket, a button hole, a pattern, a weave orientation, and a ditch formed between the two pieces of fabric joined together to form the workpiece and a feature classification data signal related to an identity of the recognized feature;
• and
• a processor configured to:
• receive, from the object recognition neural network, an indication of the location and the shape of the recognized feature from the feature detection data signal;
• receive, from the object recognition neural network, an indication of the identity of the recognized feature from the feature classification data signal;
• and
• determine, based on the identity of the recognized feature, a spatial relationship between a stitch location and the location of the recognized feature;
• and
• control the motor, the swing motor, and the feed mechanism to form stitches in the workpiece at the stitch location.

A system for detecting sewing errors in one or more stitches of a plurality of stitches formed during a sewing operation performed by a sewing machine, the sewing machine comprising a feed mechanism for moving a workpiece on a sewing bed of the sewing machine, a sewing head arranged above the sewing bed, a needle bar extending toward the sewing bed from the sewing head to a distal end to which a needle is attached, a thread threaded through the needle, and a motor for moving the needle bar in a reciprocating motion to move the needle and the thread through the workpiece to perform the sewing operation, the system comprising:

• a camera having a field of view encompassing at least a portion of any of the workpiece, the thread, and the one or more stitches formed in the workpiece, wherein the camera generates a camera data signal related to the portions of the workpiece, the thread, and the one or more stitches formed in the workpiece in the field of view of the camera;
• and
• a processor configured to:
• receive the camera data signal from the camera;
• process the camera data signal through an object recognition neural network that is trained to detect and to classify a recognized object from the camera data signal as at least one of the workpiece and the one or more stitches formed in the workpiece, wherein the object recognition neural network generates an object detection data signal related to at least one of a position, an orientation, a size, and a topography of the recognized object and an object classification data signal related to an identity of the recognized object;
• receive, from the object recognition neural network an indication of a stitch position and a stitch orientation of one or more stitches formed in the workpiece from the object recognition data signal;
• receive, from the object recognition neural network an indication of a stitch type of each of the one or more stitches formed in the workpiece from the object classification data signal;
• and
• generate, based on the indications of the stitch position, the stitch orientation, and the stitch type of the one or more stitches, sewing error data comprising an indication that one or more of the one or more stitches includes a sewing error.

A sewing machine, comprising:

• a feed mechanism arranged in a sewing bed of the sewing machine wherein a workpiece can be placed on and be at least one of moved and rotated on the sewing bed by the feed mechanism;
• a sewing head arranged above the sewing bed, the sewing head comprising:
• a needle bar extending toward the sewing bed to a distal end to which a needle is attached, wherein a thread is threaded through the needle;
• and
• a motor for moving the needle bar in a reciprocating motion to move the needle and the thread through the workpiece to form one or more stitches in the workpiece during a sewing operation;
• a camera having a field of view encompassing at least a portion of any of the workpiece, the thread, and the one or more stitches formed in the workpiece, wherein the camera generates a camera data signal related to the portions of the workpiece, the thread, and the one or more stitches formed in the workpiece in the field of view of the camera;
• an object recognition neural network that is trained to detect and to classify a recognized object from the camera data signal as at least one of the workpiece and the one or more stitches formed in the workpiece, wherein the object recognition neural network generates an object detection data signal related to at least one of a position, an orientation, a size, and a topography of the recognized object and an object classification data signal related to an identity of the recognized object;
• and
• a processor configured to:
• receive, from the object recognition neural network an indication of a stitch position and a stitch orientation of each of a plurality of stitches formed in the workpiece from the object recognition data signal;
• receive, from the object recognition neural network an indication of a stitch type of each of the plurality of stitches formed in the workpiece from the object classification data signal;
• and
• generate, based on the indications of the stitch position, the stitch orientation, and the stitch type of the plurality of stitches, sewing error data comprising an indication that one or more of the plurality of stitches includes a sewing error.