MINING LOCATIONS AND OTHER CONTEXT INFORMATION FROM CONSTRUCTION DOCUMENTS

• US 20200133970A1
• Filed: 10/30/2018
• Published: 04/30/2020
• Est. Priority Date: 10/30/2018
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: KHABIRI, Elham, Mazzoleni, Pietro
• Granted Patent: US 12,008,028 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 16/367 Ontology
• CPC Class Codes: G06F 16/387 using geographical or spatial information, e.g. location
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/047 Probabilistic or stochastic networks
• CPC Class Codes: G06N 3/088 Non-supervised learning, e.g. competitive learning
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06N 5/046 Forward inferencing; Production systems
• CPC Class Codes: G06V 30/422 Technical drawings; Geographical maps

Abstract

A method of processing construction documentation includes receiving, as input data into a processor of a computer, a document directed to a specific construction project as input data. The processor extracts location and context information from the document. The location information defines locations relevant to the specific construction project and the context information comprises text extracted from image data of the document.

Claims

1. A method of processing construction documentation, the method comprising:

receiving, as input data into a processor of a computer, a document directed to a specific construction project as input data; and

extracting, using the processor, location information and context information from the document,

wherein the location information defines locations relevant to the specific construction project and the context information comprises text extracted from image data of the document.

2. The method of claim 1, further comprising using at least one industrial standard to develop an ontology and a taxonomy for all documents directed to the specific construction project,

wherein the ontology comprises a set of concepts and categories that shows properties and interrelationships in a domain of said documents and the taxonomy comprises a common vocabulary to be used for terms found in all said documents and the extracted context information is semantically related to the taxonomy of the domain.

3. The method of claim 1, further comprising preliminarily training at least one learning model to be used in the extracting of location information and context information.

4. The method of claim 3, wherein the learning model comprises at least one of:

a convolutional neural network (CNN);

a regional-based convolution network; and

an end-to-end (E2E) analytic.

5. The method of claim 1, further comprising applying a natural language processing (NPL) to the extracted location information to semantically correlate the context information with different contexts related to the specific construction project.

6. The method of claim 1, further comprising applying an optical character recognition (OCR) processing for the extracting of location information and context information.

7. The method of claim 1, further comprising storing the document in a database of construction documents directed to the specific construction project, as annotated with the extracted location and context information.

8. The method of claim 7, wherein the received document comprises one of a plurality of documents directed to the specific construction project and wherein all documents directed to the specific construction project have location and context information extracted and are stored in the database of construction documents directed to the specific construction project, as annotated with their respective location and context information.

9. The method of claim 8, further comprising providing an interface with natural language processing (NLP) to permit users to access the database of construction documents directed to the specific construction project using input queries constructed from user inputs in a natural language.

10. The method of claim 9, as implemented in a cloud service.

11. The method of claim 1, as implemented on an analytics service platform.

12. The method of claim 1, wherein the extracting of location and context information comprises inferring physical locations and context information from engineering design documents where such information is not available either as metadata or directly accessible from a three-dimensional (3D) model of the construction project.

13. The method of claim 12, further comprising:

receiving input data as a data-dump of one or more 3D models of the construction project and corresponding architecture and an overall layout for the construction project; and

providing information from the input data as an input into an analytical model which will automatically infer grid coordinates in data directed to the construction project even when grid coordinates in an engineering document are obscured by other content in an engineering document.

14. A method of processing construction documentation, the method comprising:

using at least one industrial standard to develop an ontology and a taxonomy as defining a domain of construction documents to be used for processing all documents directed to a specific construction project, the ontology comprising a set of concepts and categories in the domain that shows their properties and interrelationships and the taxonomy comprising a common vocabulary to be used for terms found in all the documents directed to the construction project;

training a learning model, using a processor on a computer, to extract location information and context information from documents, using the developed ontology and taxonomy such that extracted text comprising the context information is semantically related to said taxonomy;

applying the trained learning model to all documents directed to the construction project; and

upon completion of applying the trained learning model, storing each document in a database of documents directed to the construction project, each document being stored as annotated with its extracted location and context information.

15. The method of claim 14, further comprising:

receiving input data as a data-dump of one or more 3D models of the construction project and corresponding architecture and an overall layout for the construction project; and

providing information from the input data as an input into an analytical model which will automatically infer grid coordinates in data directed to the construction project even when grid coordinates in an engineering document are obscured by other content in an engineering document.

16. The method of claim 15, further comprising providing an interface providing access to the database of documents directed to the construction project, the interface comprising a natural language processing (NLP) that permits a user to query the database using a natural language.

17. The method of claim 16, as implemented as a cloud service.

18. A method of processing construction documentation, the method comprising:

receiving input data as a data-dump of one or more 3D models of a construction project and corresponding architecture and an overall layout for the construction project;

providing information from the input data as an input into an analytical model which will automatically infer grid coordinates in data directed to the construction project even when grid coordinates in an engineering document are obscured by other content in an engineering document.

19. The method of claim 18, further comprising:

using at least one industrial standard to develop an ontology and a taxonomy as defining a domain of construction documents to be used for processing all documents directed to the construction project, the ontology comprising a set of concepts and categories in the domain that shows their properties and interrelationships and the taxonomy comprising a common vocabulary to be used for terms found in all the documents directed to the construction project;

training a learning model, using a processor on a computer, to extract location information and context information from documents, using the developed ontology and taxonomy such that extracted text comprising the context information is semantically related to said taxonomy;

applying the trained learning model to all documents directed to the construction project; and

upon completion of applying the trained learning model, storing each document in a database of documents directed to the construction project, each document being stored as annotated with its extracted location and context information.

20. The method of claim 19, further comprising providing an interface providing access to the database of documents directed to the construction project, the interface comprising a natural language processing (NLP) that permits a user to query the database using a natural language.