Controlling mobile construction equipment for subterranean mapping
Images, which are captured by an image capture device on a ground disturbing work machine, and that are taken from different perspectives, and are received. A machine learned image identification model identifies items of interest in the images. A three-dimensional representation is generated based on the set of images. The three-dimensional representation identifies a depth at which the recognized items lie beneath the surface of the soil being excavated. A map request is received which identifies a location and depth for which image data is to be provided. A three-dimensional representation of the location and depth are provided in response to the request.
1. A work machine computing system, comprising:
an image transmission handling system that receives a first set of images of a first excavation site that is located at a first location and a second set of images from a second excavation site at a second location;
automatic item identifier logic that recognizes a first set of items in the first excavation site from the first set of images, the first set of items being located below a surface level of ground at the first excavation site by a first subterranean depth and a second set of items in the second excavation site from the second set of images, the second set of items being located below a surface level of ground at the second excavation site by a second subterranean depth;
three-dimensional (3D) generation logic that generates a first 3D representation of the first excavation site from the first set of images, with the recognized first set of items located within the first 3D representation at the first subterranean depth based on the first set of images and a second 3D representation of the second excavation site from the second set of images, with the recognized second set of items located within the second 3D representation at the second subterranean depth based on the second set of images;
image correlation logic configured to identify correlations between the first set of items and the second set of items based on the first and second 3D representations, wherein the 3D generation logic generates a third 3D representation corresponding to a third location based on the correlations identified between the first set of items and the second set of items;
data store interaction logic that stores the first, second, and third 3D representation in a data store; and
a request processing system that receives a request, from a requesting system, identifying a requested location, accesses the data store to obtain the 3D representation corresponding to the requested location and generates a response that includes the 3D representation corresponding to the requested location and sends the response to the requesting system.
2. The work machine computing system of claim 1 wherein the 3D generation logic is configured to identify a pose, in 3D space, of each recognized item based on the first set of images and the second set of images.
3. The work machine computing system of claim 1 wherein the 3D generation logic is configured to generate the third 3D representation with recognized items located within the third 3D representation at a third subterranean depth based on the correlations identified between the first set of items and the second set of items.
4. The work machine computing system of claim 3 wherein the image correlation logic is configured to identify the correlations as extrapolation of poses of the recognized items based on the first and second 3D representations.
5. The work machine computing system of claim 2 wherein the automatic item identifier logic comprises a machine learning classification model configured to recognize the items by classifying the items into one of a plurality of predefined classes.
6. The work machine computing system of claim 2 wherein the automatic item identifier logic is configured to identify a material that a recognized item is made from.
7. The work machine computing system of claim 2 wherein the first excavation site includes a first excavation with a sidewall and wherein the automatic item identifier logic is configured to recognize a soil type along a side wall of the excavation.
8. The work machine computing system of claim 7 wherein the 3D generation logic is configured to identify different depth levels, in the 3D space, of each soil type recognized along the sidewall, based on the first set of images.
9. The work machine computing system of claim 2 wherein the request includes a depth identifier and wherein the request processing system comprises:
request parsing logic that parses the request to identify the requested location and the depth identified by the depth identifier; and
a response generator configured to generate the response that includes a 3D representation corresponding to the requested location and the depth.
10. A computer in method of controlling, an image processing system, the computer implemented method comprising:
receiving a first set of images of a first excavation site that is located at a first location;
receiving a second set of images of a second excavation site that is located at a second location;
recognizing a first set of items in the first excavation site from the first set of images, the items being located below a surface level of ground at the first excavation site by a first subterranean depth;
recognizing a second set of items in the second excavation site from the second set of images, the items being located below a surface level of ground at the second excavation site by a second subterranean depth;
generating a first 3D representation of the first excavation site from the first set of images, with the recognized first set of items located within the first 3D representation at the first subterranean depth based on the first set of images;
generating a second 3D representation of the second excavation site from the second set of images, with the recognized second set of items located within the second 3D representation at the second subterranean depth based on the second set of images;
identifying correlations between the first set of items and the second set of items based on the first and second 3D representations;
generating a third 3D representation corresponding to a third location based on the correlations identified between the first set of items and the second set of items;
storing the first, second, and third 3D representations in a data store;
receiving a request, from a requesting system, identifying a requested location;
accessing the data store to obtain the 3D representation corresponding to the requested location;
generating a response that includes the 3D representation corresponding to the requested location; and
sending the response to the requesting system.
11. The computer implemented method of claim 10 wherein generating the first 3D representation comprises identifying a pose, in 3D space, of each recognized item based on the first set of images.
12. The computer implemented method of claim 10 wherein generating the third the 3D representation comprises generating the third 3D representation with recognized items located within the third 3D representation at a third subterranean depth based on the correlations identified between the first set of items and the second set of items.