Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT

Wilder Nina; William Condori; Vicente Machaca; Juan Villegas; Eveling Castro

doi:10.1007/978-3-030-39442-4_49

Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT

Wilder Nina, William Condori, Vicente Machaca, Juan Villegas, Eveling Castro

Resultado de la investigación: Capítulo del libro/informe/acta de congreso › Contribución a la conferencia › revisión exhaustiva

5 Citas (Scopus)

Resumen

Actually, the use of deep learning in object detection gives good results, but this performance decreases when there are small objects in the image. In this work, is presented a comparison between the last version of You Only Look Once (YOLO) and You Only Look Twice (YOLT) on the problem of detecting small objects (ships) on optical satellite imagery. Two datasets were used: High-Resolution Ship Collection (HRSC) and Mini Ship Data Set (MSDS), the last one was built by us. The mean object’s width for HRSC and MSDS are 150 and 50 pixels, respectively. The results showed that YOLT is good only for small objects with 76,06% of Average Precision (AP), meanwhile, YOLO reached 69,80% in the MSDS dataset. Moreover, in the case of the HRSC dataset where have objects of different sizes, YOLT obtained a 40% of AP against 75% of YOLO.

Idioma original	Inglés
Título de la publicación alojada	Advances in Information and Communication - Proceedings of the 2020 Future of Information and Communication Conference FICC
Editores	Kohei Arai, Supriya Kapoor, Rahul Bhatia
Editorial	Springer
Páginas	664-677
Número de páginas	14
ISBN (versión impresa)	9783030394417
DOI	https://doi.org/10.1007/978-3-030-39442-4_49
Estado	Publicada - 2020
Evento	Future of Information and Communication Conference, FICC 2020 - San Francisco, Estados Unidos Duración: 5 mar. 2020 → 6 mar. 2020

Serie de la publicación

Nombre	Advances in Intelligent Systems and Computing
Volumen	1130 AISC
ISSN (versión impresa)	2194-5357
ISSN (versión digital)	2194-5365

Conferencia

Conferencia	Future of Information and Communication Conference, FICC 2020
País/Territorio	Estados Unidos
Ciudad	San Francisco
Período	5/03/20 → 6/03/20

Nota bibliográfica

Publisher Copyright:
© 2020, Springer Nature Switzerland AG.

Acceder al documento

10.1007/978-3-030-39442-4_49

Otros archivos y enlaces

Enlace a la publicación en Scopus

Citar esto

Nina, W., Condori, W., Machaca, V., Villegas, J., & Castro, E. (2020). Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT. En K. Arai, S. Kapoor, & R. Bhatia (Eds.), Advances in Information and Communication - Proceedings of the 2020 Future of Information and Communication Conference FICC (pp. 664-677). (Advances in Intelligent Systems and Computing; Vol. 1130 AISC). Springer. https://doi.org/10.1007/978-3-030-39442-4_49

Nina, Wilder ; Condori, William ; Machaca, Vicente et al. / Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT. Advances in Information and Communication - Proceedings of the 2020 Future of Information and Communication Conference FICC. editor / Kohei Arai ; Supriya Kapoor ; Rahul Bhatia. Springer, 2020. pp. 664-677 (Advances in Intelligent Systems and Computing).

@inproceedings{3684f674e123409d88377b13156b295c,

title = "Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT",

abstract = "Actually, the use of deep learning in object detection gives good results, but this performance decreases when there are small objects in the image. In this work, is presented a comparison between the last version of You Only Look Once (YOLO) and You Only Look Twice (YOLT) on the problem of detecting small objects (ships) on optical satellite imagery. Two datasets were used: High-Resolution Ship Collection (HRSC) and Mini Ship Data Set (MSDS), the last one was built by us. The mean object{\textquoteright}s width for HRSC and MSDS are 150 and 50 pixels, respectively. The results showed that YOLT is good only for small objects with 76,06% of Average Precision (AP), meanwhile, YOLO reached 69,80% in the MSDS dataset. Moreover, in the case of the HRSC dataset where have objects of different sizes, YOLT obtained a 40% of AP against 75% of YOLO.",

keywords = "Object detection, Satellite Imagery, Ship detection, Small objects, YOLO, YOLT",

author = "Wilder Nina and William Condori and Vicente Machaca and Juan Villegas and Eveling Castro",

note = "Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG.; null ; Conference date: 05-03-2020 Through 06-03-2020",

year = "2020",

doi = "10.1007/978-3-030-39442-4_49",

language = "Ingl{\'e}s",

isbn = "9783030394417",

series = "Advances in Intelligent Systems and Computing",

publisher = "Springer",

pages = "664--677",

editor = "Kohei Arai and Supriya Kapoor and Rahul Bhatia",

booktitle = "Advances in Information and Communication - Proceedings of the 2020 Future of Information and Communication Conference FICC",

}

Nina, W, Condori, W, Machaca, V , Villegas, J & Castro, E 2020, Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT. En K Arai, S Kapoor & R Bhatia (eds.), Advances in Information and Communication - Proceedings of the 2020 Future of Information and Communication Conference FICC. Advances in Intelligent Systems and Computing, vol. 1130 AISC, Springer, pp. 664-677, Future of Information and Communication Conference, FICC 2020, San Francisco, Estados Unidos, 5/03/20. https://doi.org/10.1007/978-3-030-39442-4_49

Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT. / Nina, Wilder; Condori, William; Machaca, Vicente et al.

Advances in Information and Communication - Proceedings of the 2020 Future of Information and Communication Conference FICC. ed. / Kohei Arai; Supriya Kapoor; Rahul Bhatia. Springer, 2020. p. 664-677 (Advances in Intelligent Systems and Computing; Vol. 1130 AISC).

Resultado de la investigación: Capítulo del libro/informe/acta de congreso › Contribución a la conferencia › revisión exhaustiva

TY - GEN

T1 - Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT

AU - Nina, Wilder

AU - Condori, William

AU - Machaca, Vicente

AU - Villegas, Juan

AU - Castro, Eveling

PY - 2020

Y1 - 2020

N2 - Actually, the use of deep learning in object detection gives good results, but this performance decreases when there are small objects in the image. In this work, is presented a comparison between the last version of You Only Look Once (YOLO) and You Only Look Twice (YOLT) on the problem of detecting small objects (ships) on optical satellite imagery. Two datasets were used: High-Resolution Ship Collection (HRSC) and Mini Ship Data Set (MSDS), the last one was built by us. The mean object’s width for HRSC and MSDS are 150 and 50 pixels, respectively. The results showed that YOLT is good only for small objects with 76,06% of Average Precision (AP), meanwhile, YOLO reached 69,80% in the MSDS dataset. Moreover, in the case of the HRSC dataset where have objects of different sizes, YOLT obtained a 40% of AP against 75% of YOLO.

AB - Actually, the use of deep learning in object detection gives good results, but this performance decreases when there are small objects in the image. In this work, is presented a comparison between the last version of You Only Look Once (YOLO) and You Only Look Twice (YOLT) on the problem of detecting small objects (ships) on optical satellite imagery. Two datasets were used: High-Resolution Ship Collection (HRSC) and Mini Ship Data Set (MSDS), the last one was built by us. The mean object’s width for HRSC and MSDS are 150 and 50 pixels, respectively. The results showed that YOLT is good only for small objects with 76,06% of Average Precision (AP), meanwhile, YOLO reached 69,80% in the MSDS dataset. Moreover, in the case of the HRSC dataset where have objects of different sizes, YOLT obtained a 40% of AP against 75% of YOLO.

KW - Object detection

KW - Satellite Imagery

KW - Ship detection

KW - Small objects

KW - YOLO

KW - YOLT

UR - http://www.scopus.com/inward/record.url?scp=85081377560&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-39442-4_49

DO - 10.1007/978-3-030-39442-4_49

M3 - Contribución a la conferencia

AN - SCOPUS:85081377560

SN - 9783030394417

T3 - Advances in Intelligent Systems and Computing

SP - 664

EP - 677

BT - Advances in Information and Communication - Proceedings of the 2020 Future of Information and Communication Conference FICC

A2 - Arai, Kohei

A2 - Kapoor, Supriya

A2 - Bhatia, Rahul

PB - Springer

Y2 - 5 March 2020 through 6 March 2020

ER -

Nina W, Condori W, Machaca V , Villegas J , Castro E. Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT. En Arai K, Kapoor S, Bhatia R, editores, Advances in Information and Communication - Proceedings of the 2020 Future of Information and Communication Conference FICC. Springer. 2020. p. 664-677. (Advances in Intelligent Systems and Computing). https://doi.org/10.1007/978-3-030-39442-4_49

Small Ship Detection on Optical Satellite Imagery with YOLO and YOLT

Resumen

Serie de la publicación

Conferencia

Nota bibliográfica

Acceder al documento

Otros archivos y enlaces

Huella

Citar esto