Visual tracking of objects for unmanned surface vehicle navigation

Chae, Kyung Hwa; Moon, Yong Seon; Ko, Nak Yong

doi:10.1109/iccas.2016.7832338

Cited by 5 publications

(4 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, they utilized Deep Neural Networks (DNNs) and Region Proposal Networks (RPNs) to obtain a 2D bounding box of target ships. Furthermore, a fast detection method was designed for surface objects based on ResNet (Chae et al, 2017), and the speed of object detection can reach 32.4 frames per second (FPS). Moreover, Qin (Qin and Zhang, 2018) adopted FCN for surface obstacle detection, which has a good robustness.…”

Section: Methodsmentioning

confidence: 99%

An Image-Based Benchmark Dataset and a Novel Object Detector for Water Surface Object Detection

Zhou

Sun

et al. 2021

Front. Neurorobot.

View full text Add to dashboard Cite

Water surface object detection is one of the most significant tasks in autonomous driving and water surface vision applications. To date, existing public large-scale datasets collected from websites do not focus on specific scenarios. As a characteristic of these datasets, the quantity of the images and instances is also still at a low level. To accelerate the development of water surface autonomous driving, this paper proposes a large-scale, high-quality annotated benchmark dataset, named Water Surface Object Detection Dataset (WSODD), to benchmark different water surface object detection algorithms. The proposed dataset consists of 7,467 water surface images in different water environments, climate conditions, and shooting times. In addition, the dataset comprises a total of 14 common object categories and 21,911 instances. Simultaneously, more specific scenarios are focused on in WSODD. In order to find a straightforward architecture to provide good performance on WSODD, a new object detector, named CRB-Net, is proposed to serve as a baseline. In experiments, CRB-Net was compared with 16 state-of-the-art object detection methods and outperformed all of them in terms of detection precision. In this paper, we further discuss the effect of the dataset diversity (e.g., instance size, lighting conditions), training set size, and dataset details (e.g., method of categorization). Cross-dataset validation shows that WSODD significantly outperforms other relevant datasets and that the adaptability of CRB-Net is excellent.

show abstract

Section: Methodsmentioning

confidence: 99%

An Image-Based Benchmark Dataset and a Novel Object Detector for Water Surface Object Detection

Zhou

Sun

et al. 2021

Front. Neurorobot.

View full text Add to dashboard Cite

show abstract

“…These methods have significant limitations and poor generalization because of the inclusion of artificially designed features. The emergence of deep learning has led to faster development in this field, for example, Chae et al proposed fast surface object detection based on ResNet backbone network [4] . Chen et al proposed a small vessel detection algorithm inspired by YOLOv2 [5] .…”

Section: Introductionmentioning

confidence: 99%

Improved loss and feature fusion for water surface object detection

Zhu

2023

International Conference on Image, Signal Processing, and Pattern Recognition (ISPP 2023)

View full text Add to dashboard Cite

Water surface object detection is one of the essential techniques for carrying out water-related tasks and assisting in the conduct of vessels. However, water surface object detection algorithms currently have low-precision problems detecting small and irregular objects, and difficult samples are hard to detect in harsh environments. To this end, an improved loss and enhancement feature fusion method is proposed to optimize the YOLOX object detection algorithm for water surface target detection. YOLOX, as an anchorless frame algorithm, has better detection capability for irregular objects. Improved loss improves the focus on difficult samples and treats positive samples asymmetrically based on EIoU loss. The Lite-RFP enhanced feature fusion mechanism enables the network to recursively pass contextual information recursively. It enables the shallow network to integrate deep semantic information better, improving the performance of small target detection and maintaining a lighter network structure. The experimental results show that the improved algorithm based on YOLOXs improves the mAP value by 4.11 percentage points compared to the original YOLOX-l. At the same time, a real-time detection speed of 56.534 FPS can be achieved, and the detection problems caused by different water surface environments are improved.

show abstract

“…Binocular vision technology has the advantages of fast measurement speed, high measurement accuracy, a simple structure, being noncontact, etc. It is widely used in many fields such as mechanical manufacturing [1], industrial robots [2], auto navigation [3], and 3D reconstruction [4]. Before the measurement, the parameters of the binocular vision need to be calibrated, including the intrinsic parameters of each camera and the extrinsic parameters (structure parameters) of the two cameras.…”

Section: Introductionmentioning

confidence: 99%

A simple and precise calibration method for binocular vision

Zhang

et al. 2022

Meas. Sci. Technol.

View full text Add to dashboard Cite

The binocular vision is an important part of machine vision measurement. The calibration accuracy is crucial for binocular vision. As for the determination of the structure parameters of the two cameras, the existing approaches usually obtain the initial values and optimize them according to the image-space errors, object-space errors or combination of them. In the optimization process, constructing the objective function only through the image-space errors or object-space errors is not enough. Moreover, the image-space errors and object-space errors can form a variety of combinations for constructing objective function. Therefore, it is hard to choose the error criterion. The inadequate error criterion may lead to over optimized or local minima (ambiguity solution). To solve the problem above, this paper proposes a simple and precise calibration method for binocular vision based on the points distance constraints and image-space errors. The process of determining the structure parameters was divided into non-iterative and iterative parts. We calculated the structure parameters of the two cameras according to distance constraints of every two feature points non-iteratively. The results obtained in this step were set as the initial value and refined through minimizing the reprojection errors by Levenberg-Marquardt method. Because the results obtained in the non-iterative step are accurate enough, only one iteration is needed. In this way, we finish the calibration avoiding to choose the error criterions. Furthermore, our method reduces the number of iterations to improve the calibration efficiency on the premise of guaranteeing the calibration accuracy. The experimental results show the superiority of this calibration method compared with other calibration methods.

show abstract

Visual tracking of objects for unmanned surface vehicle navigation

Cited by 5 publications

References 3 publications

An Image-Based Benchmark Dataset and a Novel Object Detector for Water Surface Object Detection

An Image-Based Benchmark Dataset and a Novel Object Detector for Water Surface Object Detection

Improved loss and feature fusion for water surface object detection

A simple and precise calibration method for binocular vision

Contact Info

Product

Resources

About