A small object detection method based on CNN in serial multispectral images

Yu Jia; Bo Lei; Chensheng Wang; Weichao Du

doi:10.1117/12.2587648

12 March 2021 A small object detection method based on CNN in serial multispectral images

Yu Jia, Bo Lei, Chensheng Wang, Weichao Du

Proceedings Volume 11763, Seventh Symposium on Novel Photoelectronic Detection Technology and Applications; 1176394 (2021) https://doi.org/10.1117/12.2587648
Event: Seventh Symposium on Novel Photoelectronic Detection Technology and Application 2020, 2020, Kunming, China

Abstract

This paper introduces a small object detection method based on 3D CNN in multispectral serial images. The temporal dimension is set as 5, which means the convolutional filter is n*n*c*5, because 5 consecutive images are sufficient to remove noises. We proposed a method for corresponding training sample preparation, where the true small object region is calculated and object moving in every direction is simulated which is called object simulation base region. The negative sample set is composed of random multispectral images while the positive sample set is composed by linear superposition of object simulation base regions. Experimental results proved this method is feasible and effective.

Citation Download Citation

Yu Jia, Bo Lei, Chensheng Wang, and Weichao Du "A small object detection method based on CNN in serial multispectral images", Proc. SPIE 11763, Seventh Symposium on Novel Photoelectronic Detection Technology and Applications, 1176394 (12 March 2021); https://doi.org/10.1117/12.2587648

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available