A method and apparatus for assembling components of a workpiece. Some embodiments include a work surface; a first digital projector that projects an image of at least some features of the workpiece onto the work surface, wherein the image includes a plurality of line indicators that have visually discernible different markings such as colors or line types; a camera that obtains input images of the work surface and the components of the workpiece; and a first controller configured to receive the input images from the camera and to control the first digital projector to project the first output digital image on the work surface for assembly of the components to form the workpiece, wherein the first controller is further configured to store distortion-correction parameters and to use the stored distortion-correction parameters to adjust the first projected output digital image. In some embodiments, the workpiece is a lumber truss.

A method of board lumber grading is performed in an industrial environment on a machine learning framework configured as an interface to a machine learning-based deep convolutional network that is trained end-to-end, pixels-to-pixels on semantic segmentation. The method uses deep learning techniques that are applied to semantic segmentation to delineate board lumber characteristics, including their sizes and boundaries.

The invention relates to a method for illuminating a wood fibre web for deviation detection. The method comprises turning on LEDs of an overhead sidelight that are side-directed towards a first edge of the web for illuminating a first half of width of the web, capturing an image of the first half, turning off the LEDs, turning on LEDs of the overhead sidelight that are side-directed towards a second edge of the web for illuminating a second half of width of the web, capturing an image of the second half, and turning off the LEDs. The invention also relates to an overhead sidelight, a machine vision system, and a lighting system.

The present disclosure provides embodiments of methods, systems, and apparatuses for detecting wet spots on machined surfaces of wood workpieces. Images of laser spots on a workpiece may be processed to determine area and aspect ratio values of the laser spots. Wet spots may be detected on the workpiece based at least on the area and aspect ratio values, and optionally based in part on color image data. A facility may use wet spot detection in grade determination and/or to classify wood pieces as wet or dry for the determination of appropriate drying conditions.

Near InfraRed NIR technology, including NIR cameras and detectors, is used to accurately identify moisture content and the specific locations of the moisture on an entire surface of a veneer sheet or other wood product. Based on the identified moisture content for a given wood product, one or more actions are taken with respect to wood product to ensure the wood product is put the most efficient, effective, and valuable use.

Near InfraRed NIR technology, including NIR cameras and detectors, and machine learning methods and systems, including one or more Machine Learning (ML) based moisture level detection models, are used to accurately identify moisture content and the specific locations of the moisture on an entire surface of a veneer sheet or other wood product and provide moisture level prediction data for the veneer sheet or other wood product. Based on the moisture level prediction data for a given wood product, one or more actions are taken with respect to wood product to ensure the wood product is put to the most efficient, effective, and valuable use.

Near InfraRed NIR technology, including NIR cameras and detectors, and one or more visual cameras are used to generate superimposed image data representing a visual/NIR composite image of wood product and correlating moisture levels with physical features of the surfaces of the wood product. By analyzing the visual/NIR composite image represented by superimposed image data, moisture pockets near an open physical feature that, absent the presence open physical feature would be a problem, can be identified and ignored. Based on the identified moisture levels at various locations in a given wood product, and the proximity of physical features to the identified moisture locations, one or more actions are taken with respect to wood product to ensure the wood product is put to the most efficient, effective, and valuable use.

Near InfraRed NIR technology, including NIR cameras and detectors, is used to accurately identify surface irregularities on a surface of a veneer sheet or other wood product. Based on the identified surface irregularities for a given wood product, one or more actions are taken with respect to wood product or production process to ensure the wood product is put to the most efficient, effective, and valuable use.

Near InfraRed NIR technology, including NIR cameras and detectors, and machine learning methods and systems, including one or more Machine Learning (ML) based surface irregularity prediction models, are used to accurately identify surface irregularities on a surface of a wood product, such as a veneer sheet or ribbon, and provide irregularity prediction data for the wood product. Based on the irregularity prediction data for a given wood product, one or more actions are taken with respect to wood product or the production process to ensure the wood product is put to the most efficient, effective, and valuable use.

Near InfraRed NIR technology, including NIR cameras and detectors, are used to detect irregularity in the surface of a wood product. Based on the detected irregularities at various locations in a given wood product, one or more actions are taken with respect to a production process used to produce the wood product to ensure the wood product is put to the most efficient, effective, and valuable use.