基于改进的YOLOv5的轻量级单阶段船舶对象检测算法用于无人水面船只.
Hui Sun1, Weizhe Zhang1, Shu Yang1
1State Key Laboratory on Integrated Optoelectronics, College of Electronic Science and Engineering, Jilin University, Changchun 130012, China.
Sensors (Basel, Switzerland)
|September 14, 2024
概括
本研究引入了船舶改进的物体检测网络,在资源有限的设备上平衡准确性和效率. 增强的YOLOv5模型显著降低参数和计算负载,同时提高检测准确性.
科学领域:
- 计算机视觉 计算机视觉
- 机器学习 机器学习
- 深度学习 (Deep Learning) 是一种深度学习.
背景情况:
- 对象检测在各种行业中至关重要,包括海上应用.
- 在资源有限的设备上部署物体检测,如无人水面船,需要高效的模型.
- 现有的模型往往难以平衡精度与较低的计算负载和参数数量.
研究的目的:
- 开发一个高效的物体检测网络,用于在具有有限计算资源的设备上执行船舶检测任务.
- 提高在海上环境中检测小物体的准确性.
- 为了减少对象检测模型的参数数量和计算负载,而不会影响性能.
主要方法:
- 一个增强的ShuffleNetV2网络被用作基于YOLOv5的对象检测模型的骨干.
- 在小物体检测层中集成了一种新的分割DLKA模块,以提高准确性.
- 使用WIOUv3损失函数来减轻低质量的数据样本的影响.
主要成果:
- 与YOLOv5s.相比,提出的方法实现了参数的71%降低和计算负载的58%降低.
- 在 IoU 值 0.5 (mAP@0.5) 的平均平均精度 (mAP) 增加了 3.9%.
- 在从0.5到0.95 (mAP@0.5:0.95) 的不同IOU值下,mAP提高了3.3%.
结论:
- 开发的物体检测网络为船舶检测的效率和准确性提供了显著的改进.
- 该模型非常适合在具有有限计算能力的边缘设备上的实时应用.
- 这项研究有助于推进海上监视和自主导航的高效深度学习模型.
相关概念视频
01:22Force Classification
1.2K
Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
1.2K
01:05Difference from Background: Limit of Detection
6.0K
The limit of detection (LOD) is the smallest amount of analyte that can be distinguished from the background noise. The LOD value corresponds to the concentration at which the analyte signal is three times larger than the standard deviation of the blank signal. Below this value, the analyte signal cannot be differentiated from the background noise. It is calculated by dividing the calibration slope by 3 times the standard deviation of the blank signals.
The LOD indicates the presence or absence...
The LOD indicates the presence or absence...
6.0K
01:31Classification of Systems-II
137
Continuous-time systems have continuous input and output signals, with time measured continuously. These systems are generally defined by differential or algebraic equations. For instance, in an RC circuit, the relationship between input and output voltage is expressed through a differential equation derived from Ohm's law and the capacitor relation,
137
01:26Classification of Systems-I
177
Linearity is a system property characterized by a direct input-output relationship, combining homogeneity and additivity.
Homogeneity dictates that if an input x(t) is multiplied by a constant c, the output y(t) is multiplied by the same constant. Mathematically, this is expressed as:
Homogeneity dictates that if an input x(t) is multiplied by a constant c, the output y(t) is multiplied by the same constant. Mathematically, this is expressed as:
177
01:30Types of Global Positioning System Surveys
51
GPS surveying methods vary in application, accuracy, and data collection techniques, catering to diverse surveying and mapping needs. Static GPS, kinematic GPS, and real-time kinematic (RTK) surveying are widely used. Each technique offers distinct advantages.Static GPS involves placing one receiver at a known reference point and another at the target point. It collects exact positional data by observing multiple satellite ranges over an extended period, achieving centimeter-level accuracy for...
51
02:07Improving Translational Accuracy
9.4K
Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...
9.4K