Abstract: The techniques described herein relate to methods, apparatus, and computer readable media configured to determine an estimated volume of an object captured by a three-dimensional (3D) point cloud. A 3D point cloud comprising a plurality of 3D points and a reference plane in spatial relation to the 3D point cloud is received. A 2D grid of bins is configured along the reference plane, wherein each bin of the 2D grid comprises a length and width that extends along the reference plane. For each bin of the 2D grid, a number of 3D points in the bin and a height of the bin from the reference plane is determined. An estimated volume of an object captured by the 3D point cloud based on the calculated number of 3D points in each bin and the height of each bin.

Abstract: The techniques described herein relate to methods and systems for three-dimensional (3D) image processing using deep learning model pre-trained with two-dimensional (2D) images. The techniques include transforming a 3D representation to a 2D map, which can be input to the model. The output of the model could be a defect segmentation mask, or a probability of the input belonging to a given category. The input 2D image can result from a particular transformation configuration. The techniques described herein provide for an effective and efficient method of identifying a transformation type for applications such as 3D data-based classification, anomaly detection and segmentation. The techniques described herein perform statistical analysis over a set of training samples to identify a transformation type. The statistical analysis can include histograms of pixel values.

Abstract: The techniques described herein relate to methods, apparatus, and computer readable media configured to determine an estimated volume of an object captured by a three-dimensional (3D) point cloud. A 3D point cloud comprising a plurality of 3D points and a reference plane in spatial relation to the 3D point cloud is received. A 2D grid of bins is configured along the reference plane, wherein each bin of the 2D grid comprises a length and width that extends along the reference plane. For each bin of the 2D grid, a number of 3D points in the bin and a height of the bin from the reference plane is determined. An estimated volume of an object captured by the 3D point cloud based on the calculated number of 3D points in each bin and the height of each bin.

Abstract: The techniques described herein relate to methods and systems for three-dimensional (3D) inspection using deep learning model pre-trained with two-dimensional (2D) images. The techniques include transforming a 3D representation (e.g., captured 3D point cloud, 3D profiles, meshes, voxels) to a 2D map, which can be input to a deep learning model pre-trained with 2D images. The 2D map includes elements disposed in an array. Each element includes a vector of a number of geometric features. Such a configuration enables the 2D map to be in a structure acceptable by the 2D deep learning model. The 2D deep learning model generates an output based on the 2D map and provides the output to a subsystem for generating an inspection result. The inspection result can have a 3D result such as a surface area and/or volume.

Abstract: The techniques described herein relate to methods, apparatus, and computer readable media configured to determine an estimated volume of an object captured by a three-dimensional (3D) point cloud. A 3D point cloud comprising a plurality of 3D points and a reference plane in spatial relation to the 3D point cloud is received. A 2D grid of bins is configured along the reference plane, wherein each bin of the 2D grid comprises a length and width that extends along the reference plane. For each bin of the 2D grid, a number of 3D points in the bin and a height of the bin from the reference plane is determined. An estimated volume of an object captured by the 3D point cloud based on the calculated number of 3D points in each bin and the height of each bin.

Abstract: Techniques for scanning through structured light techniques that are robust to global illumination effects and that provide an accurate determination of position as provided. According to some aspects, described techniques may be computationally efficient compared with conventional techniques by making use of a novel approach of selecting frequencies of patterns for projection onto a target that mathematically enable efficient calculations. In particular, the selected frequencies may be chosen so that there is a known relationship between the frequencies. This relationship may be derived from Chebyshev polynomials and also relates the chosen frequencies to a low frequency pattern.

Abstract: Techniques for scanning through structured light techniques that are robust to global illumination effects and that provide an accurate determination of position as provided. According to some aspects, described techniques may be computationally efficient compared with conventional techniques by making use of a novel approach of selecting frequencies of patterns for projection onto a target that mathematically enable efficient calculations. In particular, the selected frequencies may be chosen so that there is a known relationship between the frequencies. This relationship may be derived from Chebyshev polynomials and also relates the chosen frequencies to a low frequency pattern.