<abstract xmlns="http://www.w3.org/1999/xhtml"><p>In In the process of sea surface object detection by remote sensing images, this paper proposes an improved Faster R-CNN sea surface object detection algorithm for the problems of low accuracy and long detection time using deep neural networks. Firstly, the size of the bounding box in the region proposal networks is analyzed, and then the object in the image is clustered using the K-Means algorithm, and the clustering results are input into region proposal networks, thereby realizing the region proposal networks improvement of. Secondly, the Soft-NMS algorithm is used to screen the target candidate frame to obtain the detected sea objects. The experimental results show that the algorithm of this paper can detect the sea surface object of remote sensing images under complex sea conditions, and its mAP can reach 87.25%, which is an average increase of 3.75% compared with the commonly used detection methods.</p></abstract>

In In the process of sea surface object detection by remote sensing images, this paper proposes an improved Faster R-CNN sea surface object detection algorithm for the problems of low accuracy and long detection time using deep neural networks. Firstly, the size of the bounding box in the region proposal networks is analyzed, and then the object in the image is clustered using the K-Means algorithm, and the clustering results are input into region proposal networks, thereby realizing the region proposal networks improvement of. Secondly, the Soft-NMS algorithm is used to screen the target candidate frame to obtain the detected sea objects. The experimental results show that the algorithm of this paper can detect the sea surface object of remote sensing images under complex sea conditions, and its mAP can reach 87.25%, which is an average increase of 3.75% compared with the commonly used detection methods.

Improved Faster R-CNN Algorithm for Sea Object Detection Under Complex Sea Conditions

International Journal of Advanced Network, Monitoring and Controls

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

In In the process of sea surface object detection by remote sensing images, this paper proposes an improved Faster R-CNN sea surface object detection algorithm...

Parameters	Values	Parameters	Values
LEARNING_RATE	le-3	Bateh_size	256
Anchor_Scales	[8,16,32]	Anchor_RATIOS	[0.7, 1.2, 1.3, 2.9]
ITERS	85000	num_classes	6
SOFT_NMS	1

Detection method	Passenger_ ship	Cargo_ship	Container_ ship	Aircraft_ ship	War_ship	mAP
VGG-16 structure	50.9%	80.6%	92.4%	98.1%	92.8%	82.96%
ResNet101 structure	54.1%	81.0%	92.8%	99.3%	93.0%	84.04%
improved ResNet101 structure	66.9%	82.3%	93.65%	99.51%	93.9%	87.25%

number	K=1	K=2	K=3	K=4	K=5	K=6	K=7
	0.6	1.3	0.5	0.7	0.6	0.5	0.5
	-	1.4	1.2	1.2	1.2	0.7	0.6
	-	-	1.6	1.3	1.3	1.2	1.1
results	-	-		2.9	1.7	1.3	1.2
	-	-	-	-	2.6	2.4	1.8
	-	-	-	-	-	3	2.1
	-	-	-	-	-	-	2.9
mAP(%)	83.84	84.3	84.32	84.98	84.29	84.57	84.34
Time(s)	503	510	515	518	524	515	517

Improved Faster R-CNN Algorithm for Sea Object Detection Under Complex Sea Conditions

Publicado en línea: 13 jul 2020

Páginas: 76 - 82

DOI: https://doi.org/10.21307/ijanmc-2020-020

Palabras clave
Sea Surface Object Detection, Faster R-CNN Algorithm, K-Means Clustering Algorithm, Soft-NMS

© 2020 Liu Yabin et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

VALUES OF MAIN NETWORK PARAMETERS

COMPARISON OF DETECTION RESULTS OF THREE NETWORK STRUCTURES

ANCHOR.RATIOS FOR DIFFERENT K VALUES

Improved Faster R-CNN Algorithm for Sea Object Detection Under Complex Sea Conditions

Publicado en línea: 13 jul 2020

Páginas: 76 - 82

DOI: https://doi.org/10.21307/ijanmc-2020-020

Palabras claveSea Surface Object Detection, Faster R-CNN Algorithm, K-Means Clustering Algorithm, Soft-NMS

© 2020 Liu Yabin et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

VALUES OF MAIN NETWORK PARAMETERS

COMPARISON OF DETECTION RESULTS OF THREE NETWORK STRUCTURES

ANCHOR.RATIOS FOR DIFFERENT K VALUES

Palabras clave
Sea Surface Object Detection, Faster R-CNN Algorithm, K-Means Clustering Algorithm, Soft-NMS