Robust Tracker: Integrating CPM-YOLO and BOTSORT for Cross-Modal Vessel Tracking

  • 0College of Information Engineering, Shanghai Maritime University, Shanghai 201306, China.

|

|

Summary

This summary is machine-generated.

This study introduces an advanced maritime vessel tracking system using YOLOv8 and BOTSORT with re-identification, enhancing detection accuracy and robustness in challenging marine conditions. The method achieves high performance without compromising real-time processing speeds.

Area Of Science

  • Computer Vision
  • Maritime Surveillance
  • Object Tracking

Background

  • Maritime environments present complex challenges for vessel detection and tracking, including dense targets, scale variations, and occlusions.
  • Existing methods often struggle with accuracy and robustness in these demanding conditions.

Purpose Of The Study

  • To develop a high-accuracy and robust multi-object tracking method for maritime vessel detection and tracking.
  • To improve detection of small-scale vessels and enhance tracking stability and reliability in complex marine scenarios.

Main Methods

  • An enhanced YOLOv8-based detector with lightweight feature enhancement and attention mechanisms was utilized for improved detection.
  • A tracking framework combining BOTSORT with an OSNet-based re-identification (ReID) model was employed for stable vessel association.
  • The approach was evaluated on Near-Infrared On-Shore (NIR) and Visible On-Shore (VIS) datasets.

Main Results

  • The proposed method demonstrated significant improvements in Precision, Recall, mAP@0.5, and mAP@0.5:0.95 on the NIR dataset compared to baseline YOLOv8.
  • It achieved superior detection accuracy and robustness on the VIS dataset, outperforming state-of-the-art approaches.
  • The framework showed advantages in trajectory continuity, occlusion handling, and identity preservation in multi-object tracking tasks.

Conclusions

  • The developed method offers a favorable balance between detection accuracy, tracking robustness, and model efficiency for practical maritime applications.
  • It achieves high performance, including competitive inference speeds (188 FPS on NIR, 187 FPS on VIS), without sacrificing real-time capabilities.
  • The approach exhibits strong cross-modal generalization capability, proving effective across different visual conditions.

Related Concept Videos

Cross-Sectional Research 01:50

12.7K

In cross-sectional research, a researcher compares multiple segments of the population at the same time. If they were interested in people's dietary habits, the researcher might directly compare different groups of people by age. Instead of following a group of people for 20 years to see how their dietary habits changed from decade to decade, the researcher would study a group of 20-year-old individuals and compare them to a group of 30-year-old individuals and a group of 40-year-old...

Crossing Over 01:34

172.3K

Unlike mitosis, meiosis aims for genetic diversity in its creation of haploid gametes. Dividing germ cells first begin this process in prophase I, where each chromosome—replicated in S phase—is now composed of two sister chromatids (identical copies) joined centrally.
The homologous pairs of sister chromosomes—one from the maternal and one from the paternal genome—then begin to align alongside each other lengthwise, matching corresponding DNA positions in a process...

Crossing Over 01:30

6.6K

Crossing over is the exchange of genetic information between homologous chromosomes during prophase I of meiosis I. Genetic recombination gives rise to allelic diversity in the newly formed daughter cells. In humans, crossing over produces genetically distinct haploid egg and sperm cells that undergo fertilization to produce unique offspring. Before cell division starts, the germ cell’s chromosome(s) undergo duplication in the S phase of the cell cycle. As the cells enter prophase I,...

Sensory Modalities 01:15

4.0K

Sensation typically is the process by which the sensory receptors and sense organs detect stimuli from the internal and external environment and transmit this information to the central nervous system for processing.
General senses refer to the broad category of sensory information detected by receptors in the body and can be further grouped into somatic and visceral senses. Somatic sensations include touch, pressure, temperature, and pain and are essential for navigating our environment and...

Monohybrid Crosses 01:20

239.8K

Overview

In the 1850s and 1860s, Gregor Mendel investigated inheritance by performing monohybrid crosses in pea plants. He crossed two plants that were true-breeding for different traits. Based on his observations, Mendel proposed that organisms inherit two copies of each trait, one from each parent, and that dominant traits can hide recessive traits. These results formed the basis of two fundamental principles in genetics: the Principle of Uniformity and the Law of Segregation.

Monohybrid...

Dihybrid Crosses 01:18

81.5K

Overview

To determine whether traits are inherited together or separately, Gregor Mendel crossed pea plants that differed in two traits. These parental plants were homozygous for both traits but displayed different phenotypes. The first generation of offspring were all dihybrids, heterozygotes exhibiting the two dominant phenotypes. When self-fertilized, the dihybrids consistently produced progeny with a 9:3:3:1 ratio of four possible phenotype combinations. This ratio suggested that...