BSU Bulletin. Mathematics, Informatics
Bibliographic description:
,
,
PRE-PROCESSING OF SPACE IМAGES IN ТНЕ OBJECT SEARCH TASK // BSU Bulletin. Mathematics, Informatics. - 2018. №4. . - С. 16-21.
Title:
PRE-PROCESSING OF SPACE IМAGES IN ТНЕ OBJECT SEARCH TASK
Financing:
Codes:
Annotation:
Тhе article discusses the joint usage of structural methods and neural networks for the detection ofman-made objects. lt means that in the method ofsearching objects on user request in the SOQL language а neural network classifier is used to evaluate the spectral and texture features. lt is possiЫe to apply one classifier to different images.
Keywords:
neural networks; classification; TensorFlow; logical recognition meth ods.
List of references:
Potaturkin О. 1., Borzov S. М., Potaturkin А О., Uzilov S. В. Metody i tekhnologii obrabotki multi- i giperspektralnykh dannykh distantsionnogo zondiro vaniya Zemli vysokogo razresheniya [Methods and Technologies for Processing Multi and Hyperspectral Data ofHigh-Resolution Remote Sensing ofthe Earth]. Vychislitel nye tekhnologii - Computational Technologies. 2013. V. 18. Рр. 60-67.
Yari D., Mokhtarzade М., Ebadi Н., Ahmadi S. Automatic Reconstruction of Regular Buildings Using а Shape-Based Balloon Snake Model. Photogrammetric Re cord. 2014. V. 29, No. 146. Рр. 187-205.
Kolbe Т. Н., Plumer L., Cremers А В. Using Constraints for the ldentification of Buildings in Aerial lmages. Procedings of the 2nd lnternational Conference оп the Practical Application ofConstraint Technology. 1996. Рр. 1-12.
Bychkov 1. V., Ruzhnikov G. М., Fedorov R. К. , Avramenko Yu. V. lnterpreta tor yazyka SOQL dlya obrabotki rastrovykh izobrazhenii SOQL [lnterpreter for Bitmap Processing]. Vychislitelnye tekhnologii - Computational Technologies. 2016. V. 21, No. 1. Рр. 49-59.
Rashid Т. Make Your Own Neural Network. CreateSpace lndependent PuЫish ing Platform, 2016. 222 р.
Кrizhevsky А , Sutskever 1., and Нinton G. Е. lmagenet Classification with Deep Convolutional Neural Networks. Advances in Neural lnformation Processing Systems. 2012. Рр. 1097-1105.