| Predicate |
Object |
| assignee |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_1fefd66fdbc3870bbaa8e5a4f87bdb0a |
| classificationCPCAdditional |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-10004
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-20081
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-30164 |
| classificationCPCInventive |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-0006
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-62
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-73 |
| classificationIPCInventive |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-00
http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-62
http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-73 |
| filingDate |
2019-07-04-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| inventor |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_23d64d971134e05a8640470735a431bd
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_9087c2c922b9b383583eff205fa4a67d
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_10516a904ee6d9f8f61e2c276a5d5296
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_ff44b5031e3bd875dfff9d692e5ff383
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_7d40fb5e5bf43224abc475d83fee0def
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_ce03f3c847099ce19526f3255e2d3dba
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_c4a72a206d1ec329d8a2ee6bd21377c0 |
| publicationDate |
2019-11-08-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| publicationNumber |
CN-110428398-A |
| titleOfInvention |
A Defect Detection Method for High-speed Railway Catenary Cable Defects Based on Deep Learning |
| abstract |
The invention discloses a deep learning-based detection method for the defect detection of the cable of the high-speed railway catenary. Perform preprocessing and format conversion; input the converted data samples into the built network model, and output the prediction results; post-process the prediction results to obtain the final defect detection results. The present invention improves the robustness of the network to the angle of the cable by modeling the inherent attributes of the artificial object, and also improves the accuracy of the traditional positive frame target detection. At the same time, the method proposed in the present invention abandons the complicated process of manually designing features, and directly uses the powerful feature learning ability of the deep network to extract the defect features of key components, realizing end-to-end defect detection, and replacing the process of manual screening of defect images, reducing It reduces the workload of people and shortens the process from finding defects to repairing. |
| isCitedBy |
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-111798447-A
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-115294451-B
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-115294451-A |
| priorityDate |
2019-07-04-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| type |
http://data.epo.org/linked-data/def/patent/Publication |