Vision-based Object Tracking in Marine Environments using Features from Neural Network Detections

Schöller, Frederik Emil Thorsson; Blanke, Mogens; Plenge-Feidenhans’, M.K.; Nalpantidis, Lazaros

doi:10.1016/j.ifacol.2020.12.1455

Cited by 15 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Introduction: Deep neural networks provide strong tools for solving problems such as visual object detection [1,2]. However, such capabilities often come with the requirement of considerable amount of carefully annotated training examples.…”

mentioning

confidence: 99%

Reweighting neural network examples for robust object detection at sea

Becktor

Boukas

Blanke

et al. 2021

Electronics Letters

Self Cite

View full text Add to dashboard Cite

Deep neural networks have had profound significance in addressing visual object detection and classification tasks. However, though with the caveat of needing large amounts of annotated training data. Furthermore, the possibility of neural networks overfitting to the biases and faults included in their respective datasets. In this work, methods for achieving robust neural networks, able to tolerate untrusted and possibly erroneous training data, are explored. The proposed method is shown to improve performance and help neural networks learn from untrusted data, provided a thoroughly annotated subset.

show abstract

mentioning

confidence: 99%

Reweighting neural network examples for robust object detection at sea

Becktor

Boukas

Blanke

et al. 2021

Electronics Letters

Self Cite

View full text Add to dashboard Cite

show abstract

“…Detection could also be based on obstacle segmentation methods such as investigated by [34]. Tracking of the subject could be based on subsequent detection [35] or by using optical flow [36]. While recording the buoy lights, the camera was often set out of focus on purpose, as it was found to give better color reproduction.…”

Section: A Synthetic Data Generationmentioning

confidence: 99%

Buoy Light Pattern Classification for Autonomous Ship Navigation Using Recurrent Neural Networks

Schöller

Nalpantidis

Blanke

2022

IEEE Trans. Intell. Transport. Syst.

View full text Add to dashboard Cite

In near coast navigation, buoys and beacons convey essential information about dangers. At night-time, selected buoys send out individual blink-sequences that are marked in sea charts. International regulations require that navigation officer on watch makes visual confirmation of objects and their type in order to navigate safely. With rapid developments of highly automated vessels, this duty needs be carried out by algorithms that detect and locate objects without human intervention. At night-time, this requires algorithms that decode blink sequences and are able to classify from this information. The paper investigates this problem and suggests an algorithm that solves the problem. Convolutional Neural Networks (CNN) with Gated Recurrent Units (GRU) are developed for classification. A dedicated architecture is suggested that includes both temporal and color decoding to obtain unique precision. We demonstrate how networks are trained on synthetically generated data, and the paper shows, on real-world data, how the suggested approach yields 100.0% accurate results on 44 real-world recordings while being robust to inaccuracy in actual blink sequences. Comparison with baseline signal processing and with a recent state-of-the-art 3D CNN model shows that the new blink-sequence classifier outperforms alternative algorithms.A showcase of the results of this work is available in this video: https://youtu.be/KEi8qNnKV2w.

show abstract

“…The automotive sector was an early adopter of imaging sensors, finding applications for both driver assistance systems [2] and autonomy [3]. Maritime applications typically combine cameras with other sensors such as radar [4], [5] or AIS [6], however, some applications of camera-only tracking exist [7], [8].…”

Section: Introductionmentioning

confidence: 99%

Maritime Tracking With Georeferenced Multi-Camera Fusion

2023

View full text Add to dashboard Cite

Cameras form an essential part of any autonomous surface vehicle's sensor package, both for COLREGs compliance to detect light signals and for identifying and tracking other vessels. Due to limited fields of view compared to more traditional autonomy sensors such as lidars and radars, an autonomous surface vessel will typically be equipped with multiple cameras which can induce biases when used in tracking if a target is present in multiple image frames. In this work, we propose a novel pipeline for camera-based maritime tracking that combines georeferencing with clustering-based multi-camera fusion for bias-free camera measurements with target range estimates. Using real-world datasets collected using the milliAmpere research platform the performance of this pipeline exceeded a lidar benchmark across multiple performance measures, both in pure detection performance and as part of a JIPDA-based tracking system.

show abstract

Vision-based Object Tracking in Marine Environments using Features from Neural Network Detections

Cited by 15 publications

References 15 publications

Reweighting neural network examples for robust object detection at sea

Reweighting neural network examples for robust object detection at sea

Buoy Light Pattern Classification for Autonomous Ship Navigation Using Recurrent Neural Networks

Maritime Tracking With Georeferenced Multi-Camera Fusion

Contact Info

Product

Resources

About