Enhancement of Single Shot Multibox Detector for Aerial Infrared Target Detection

Jiangrong Xie; Fanming Li; Hong Wei; Bing Li; Baotai Shao

doi:10.3788/AOS201939.0615001

Acta Optica Sinica, Volume. 39, Issue 6, 0615001(2019)

Enhancement of Single Shot Multibox Detector for Aerial Infrared Target Detection

Jiangrong Xie^1,2,3, Fanming Li^1,3、*, Hong Wei¹, Bing Li¹, and Baotai Shao^1,2,3

Author Affiliations

¹ Shanghai Institute of Technical Physics, Chinese Academy of Sciences, Shanghai 200083, China

² University of Chinese Academy of Sciences, Beijing 100049, China

³ Key Laboratory of Infrared System Detection and Imaging Technology, Chinese Academy of Sciences, Shanghai 200083, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(10)

Fig. 1. Structure of SSD network

Download full size

Fig. 2. Schematics of multiple feature fusion methods. (a) Pooling; (b) transposed deconvolution; (c) bi-direction fusion

Download full size

Fig. 3. Diagram of semantic segmentation branch

Download full size

Fig. 4. Comparison of detection results of small targets obtained by original SSD and improved SSD. (a) Original SSD; (b) improved model

Download full size

Fig. 5. Detection results of infrared aerial targets

Download full size

Fig. 6. Comparison of recall-precision curve of infrared aerial targets

Download full size

Table 1. Convolution receptive field and mapping image region of default boxes of SSD_300×300

View table

Table 1. Convolution receptive field and mapping image region of default boxes of SSD_300×300

Convolutional layer	Convolutional receptive field /(pixel×pixel)	Output scale of feature layer /(pixel×pixel)	Default boxes ratio	Mapping region scale /(pixel×pixel)
conv4_3	92×92	38×38	0.10	30×30
conv7	276×276	19×19	0.20	60×60
conv8_2	340×340	10×10	0.37	111×111
conv9_2	468×468	5×5	0.54	162×162
conv10_2	724×724	3×3	0.71	213×213
conv11_2	980×980	1×1	0.88	264×264

Table 2. Small object detection results of VOC2007 dataset
View table
Table 2. Small object detection results of VOC2007 dataset
Method mAP Detection result
Aero plane Bird Boat Bottle Car Dog Sheep Person
SSD_300×300 0.537 0.601 0.525 0.426 0.374 0.720 0.556 0.538 0.563
Proposed method 0.608 0.685 0.570 0.534 0.503 0.748 0.597 0.652 0.591

Table 3. Aerial target detection results of infrared dataset
View table
Table 3. Aerial target detection results of infrared dataset
Method mAP Detection result
Fighter_J Helicopter Fighter_S Airliner Bird
SSD_300×300 0.618 0.659 0.615 0.266 0.819 0.732
YOLOv3-320 0.641 0.730 0.548 0.314 0.867 0.747
Proposed method 0.705 0.784 0.636 0.485 0.822 0.796

Table 4. Number of predictive boxes for each classi?ed network and speed comparison
View table
Table 4. Number of predictive boxes for each classi?ed network and speed comparison
Method Number of boxes Total boxes FPS
75×75 38×38 19×19 10×10 5×5 3×3 1×1
Original SSD 0 4 6 6 6 4 4 8732 25.2
Modified SSD 4 4 4 4 4 4 4 30260 9.4

Tools

Get Citation

Copy Citation Text

Jiangrong Xie, Fanming Li, Hong Wei, Bing Li, Baotai Shao. Enhancement of Single Shot Multibox Detector for Aerial Infrared Target Detection[J]. Acta Optica Sinica, 2019, 39(6): 0615001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Machine Vision

Received: Dec. 18, 2018

Accepted: Feb. 19, 2019

Published Online: Jun. 17, 2019

The Author Email: Li Fanming (lfmjws@163.com)

DOI:10.3788/AOS201939.0615001

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Convolution receptive field and mapping image region of default boxes of SSD_300×300

Table 1. Convolution receptive field and mapping image region of default boxes of SSD_300×300

Table 2. Small object detection results of VOC2007 dataset

Table 2. Small object detection results of VOC2007 dataset

Table 3. Aerial target detection results of infrared dataset

Table 3. Aerial target detection results of infrared dataset

Table 4. Number of predictive boxes for each classi?ed network and speed comparison

Table 4. Number of predictive boxes for each classi?ed network and speed comparison