Interactive visualizations for industrial inspections

1 . A computer implemented method comprising:

collecting inspection image data of a part for an industrial process associated with producing the part;

generating a 3D textured model of the part that includes inspection data collected for the industrial process associated with producing the part by mapping texture data included in the inspection image data to a 3D model of the part;

tracking a viewing perspective of a user relative to a reference part;

generating pose data based on the tracked viewing perspective;

rendering an image that includes inspection data from the 3D textured model based on the pose data;

processing the rendered image using segmentation or region-of-interest (ROI) detection based on the inspection data included in the rendered image to identify one or more regions of interest of the part;

generating a segmented inspection image that highlights the one or more identified regions of interest; and

rendering a further image for display that includes image data corresponding to the reference part overlaid with the segmented inspection image.

2 . The computer implemented method of claim 1 wherein the tracking comprises receiving video image data that includes successive images of the reference part, generating the pose data comprises mapping, based on the video image data, physical locations and orientations of the reference part to a reference coordinate system that enables the inspection data from the 3D textured model to be mapped to the reference part over the successive images, and the rendering comprises successively rendering images that correspond to the successive images of the reference part.

3 . The method of claim 2 wherein the reference part is a physical part having an actual geometry that may vary relative to geometry data of the 3D textured model of the part, and the method further comprises determining geometry variation data that enables the geometry data for the 3D textured model to be conformed to the actual geometry of the reference part, wherein successively rendering the images compromises is based on both the pose data and the geometry variation data.

4 . The method of claim 3 wherein determining the geometry variation data comprises perturbing 3D model geometry data to determine offsets for a plurality of reference points of the 3D model geometry data to corresponding points of the reference part.

5 . The method of claim 2 wherein the inspection data includes texture data corresponding to one or more of: near-infrared (NIR) image data, infrared (IR) image data, and/or visible light image data, wherein rendering the images comprises overlaying the texture data on the successive images of the reference part to provide augmented images.

6 . The method of claim 2 comprising processing the inspection data collected for the industrial process associated with producing the part to visually enhance features of interest in the rendered images.

7 . The method of claim 6 wherein processing the inspection data comprises using a machine learned prediction function to generate classification data for one or more regions of the part based on the inspection data, the classification data being overlaid on a representation of the part in the rendered images.

8 . The method of claim 1 comprising receiving a user input selecting a region of the reference part, the rendering comprising displaying an indication of the user input.

9 . The method of claim 1 wherein the inspection data includes inspection data collected from multiple industrial processes associated with producing the part, the method comprising selectively including, in the rendered image, the inspection data collected from different industrial processes, based on detected user inputs.

10 . The method of claim 1 wherein:

(i) the tracking and the rendering uses one or more of an augmented/mixed reality headset including a head-mounted digital display and camera; (ii) the tracking and the rendering uses a mobile device including a digital display and connected camera; (iii) the tracking and the rendering uses a separate camera and digital display; or (iv) the tracking and the rendering uses a camera and digital projector; and

the rendering comprises overlaying computer-generated data derived from the inspection data over physical world images of the reference part.

11 . The method of claim 1 wherein the reference part is the same physical part as the part.

12 . The method of claim 1 wherein the reference part is a proxy object for the part.

13 . An interactive inspection system comprising:

an inspection device configured to collect inspection image data of a part for an industrial process associated with producing the part;

a model generation module configured to generate a 3D textured model of the part that includes inspection data collected for the industrial process associated with producing the part by mapping texture data included in the inspection image data to a 3D model of the part;

a tracking device configured for tracking a viewing perspective of a user relative to a reference part;

a computer implemented pose estimator configured for generating pose data indicating a pose of the reference part in reliance on the tracked viewing perspective;

a computer implemented rendering unit configured for generating a rendered image of the part based on the pose data, the rendered image including inspection data from the 3D textured model;

an image enhancement module configured to process the rendered image using segmentation or region-of-interest (ROI) detection based on the inspection data included in the rendered image to identify one or more regions of interest of the part, and to generate a segmented inspection image that highlights the one or more identified regions of interest; and

a display device configured to display a further image that includes image data corresponding to the reference part overlaid with the segmented inspection image.

14 . The interactive inspection system of claim 13 wherein:

the tracking device includes a camera receiving video image data that includes successive images of the reference part;

the pose estimator generates the pose data by mapping, based on the video image data, physical locations and orientations of the reference part over the successive images to a reference coordinate system; and

the rendering unit maps inspection data from the 3D textured model to the reference part over the successive images based on the pose data and generates successive corresponding rendered images including the mapped inspection data.

15 . The interactive inspection system of claim 14 wherein the reference part is a physical part having an actual geometry that varies relative to geometry data of the 3D textured model of the part, and the system further includes a geometry variation module configured to compute geometry variation data that enables the geometry data for the 3D textured model to be conformed to the actual geometry of the reference part, wherein the rendered images are generated by the rendering unit based on both the pose data and the geometry variation data.

16 . The interactive inspection system of claim 15 wherein the geometry variation module is configured to compute the geometry variation data using a process that includes perturbing 3D model geometry data to determine offsets for a plurality of reference points of the 3D model geometry data to corresponding points of the reference part.

17 . The interactive inspection system of claim 14 wherein the inspection data includes texture data corresponding to one or more of: near-infrared (NIR) image data, infrared (IR) image data, and/or visible light image data, wherein the rendered images are augmented with the texture data overlaid on the successive images of the reference part.

18 . The interactive inspection system of claim 14 further comprising an image enhancer associated with the rendering unit, the image enhancer configured to process the inspection data collected for the industrial process associated with producing the part to visually enhance features of interest in the rendered images.

19 . The interactive inspection system of claim 18 wherein the image enhancer includes a machine learned prediction function for generating classification data for one or more regions of the part based on the inspection data, the classification data being overlaid on a representation of the part in the rendered images.

20 . The interactive inspection system of claim 13 wherein:

(i) the tracking device and the display device are implemented using an augmented reality headset that includes a head-mounted digital display and camera;

(ii) the tracking device and the display device are implemented using mobile device that includes a digital display and connected camera;

(iii) the tracking device and the display device comprise a separate camera and digital display, respectively; or

(iv) the tracking device and the display device comprise a camera and digital projector, respectively;

and

the rendered image comprises a computer-generated data derived from the inspection data overlaid onto a physical world images of the reference part.