Citation: Ge, H.; Dai, Y.; Zhu, Z.;
Zang, X. Single-Stage Underwater
Target Detection Based on Feature
Anchor Frame Double Optimization
Network. Sensors 2022, 22, 7875.
https://doi.org/10.3390/s22207875
Academic Editor: Ioannis E. Livieris
Received: 24 August 2022
Accepted: 6 October 2022
Published: 17 October 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Single-Stage Underwater Target Detection Based on Feature
Anchor Frame Double Optimization Network
Huilin Ge , Yuewei Dai *, Zhiyu Zhu * and Xu Zang
School of Maine, Jiangsu University of Science and Technology, Zhenjiang 212003, China
* Correspondence: dyw@nuist.edu.cn (Y.D.); zhuzy@just.edu.cn (Z.Z.)
Abstract:
Objective: The shallow underwater environment is complex, with problems of color
shift, uneven illumination, blurring, and distortion in the imaging process. These scenes are very
unfavorable for the reasoning of the detection network. Additionally, typical object identification
algorithms struggle to maintain high resilience in underwater environments due to picture domain
offset, making underwater object detection problematic. Methods: This paper proposes a single-
stage detection method with the double enhancement of anchor boxes and features. The feature
context relevance is improved by proposing a composite-connected backbone network. The receptive
field enhancement module is introduced to enhance the multi-scale detection capability. Finally, a
prediction refinement strategy is proposed, which refines the anchor frame and features through two
regressions, solves the problem of feature anchor frame misalignment, and improves the detection
performance of the single-stage underwater algorithm. Results: We achieved an effect of 80.2 mAP
on the Labeled Fish in the Wild dataset, which saves some computational resources and time while
still improving accuracy. On the original basis, UWNet can achieve 2.1 AP accuracy improvement
due to the powerful feature extraction function and the critical role of multi-scale functional modules.
At an input resolution of 300
×
300, UWNet can provide an accuracy of 32.4 AP. When choosing the
number of prediction layers, the accuracy of the four and six prediction layer structures is compared.
The experiments show that on the Labeled Fish in the Wild dataset, the six prediction layers are better
than the four. Conclusion: The single-stage underwater detection model UWNet proposed in this
research has a double anchor frame and feature optimization. By adding three functional modules,
the underwater detection of the single-stage detector is enhanced to address the issue that it is simple
to miss detection while detecting small underwater targets.
Keywords:
underwater object detection multi-scale; dynamic convolution; UWNet; compound
connection network
1. Introduction
Due to the complex underwater environment, the turbidity of the water body, the
absorption of light by the water body, and the high cost of underwater video acquisition,
machine vision still has much room for development in the field of aquatic biological
monitoring. Underwater robots can realize the function of allowing robots to complete
specific underwater tasks instead of manual diving. Underwater robots are widely used in
safety search and rescue, pipeline inspection, oil exploration, and fishing.
The movement and operation of underwater robots are usually remotely controlled by
operators on water ships and interact through vision and control systems [
1
]. Equipped
with sonar, laser systems, cameras, and other equipment, real-time video and sonar images
are provided by underwater robots for water operators. To grasp the target in the dark un-
derwater environment, the underwater robot will also be equipped with a mechanical arm
and a searchlight. However, it is not enough for underwater robots to achieve autonomous
target grasping with the above equipment and sensors alone, and a set of underwater target
Sensors 2022, 22, 7875. https://doi.org/10.3390/s22207875 https://www.mdpi.com/journal/sensors
