Marine target detection based on Marine-Faster R-CNN for navigation radar plane position indicator images

Chen, Xiaolong; Mu, Xiaoqian; Guan, Jian; ningbo, liu; Zhou, Wei

doi:10.1631/fitee.2000611

Cited by 14 publications

(4 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, some studies have achieved high levels of ship recognition accuracy and effective interference suppression using two-stage recognition algorithms. For instance, Chen et al [8] made several improvements to Faster R-CNN in multiple aspects, including optimizing the backbone network, sample data balancing, and scale normalization, aiming to enhance the algorithm's accuracy and robustness. Additionally, differential neural architecture search, label reassignment, and various types of feature pyramid structures have been widely applied in algorithm design, yielding promising recognition results [9][10][11].…”

Section: Ship Identification Methods Under Radar and Other Scenariosmentioning

confidence: 99%

MrisNet: Robust Ship Instance Segmentation in Challenging Marine Radar Environments

Ma,

Kang,

Chen

et al. 2023

JMSE

View full text Add to dashboard Cite

In high-traffic harbor waters, marine radar frequently encounters signal interference stemming from various obstructive elements, thereby presenting formidable obstacles in the precise identification of ships. To achieve precise pixel-level ship identification in the complex environments, a customized neural network-based ship segmentation algorithm named MrisNet is proposed. MrisNet employs a lightweight and efficient FasterYOLO network to extract features from radar images at different levels, capturing fine-grained edge information and deep semantic features of ship pixels. To address the limitation of deep features in the backbone network lacking detailed shape and structured information, an adaptive attention mechanism is introduced after the FasterYOLO network to enhance crucial ship features. To fully utilize the multi-dimensional feature outputs, MrisNet incorporates a Transformer structure to reconstruct the PANet feature fusion network, allowing for the fusion of contextual information and capturing more essential ship information and semantic correlations. In the prediction stage, MrisNet optimizes the target position loss using the EIoU function, enabling the algorithm to adapt to ship position deviations and size variations, thereby improving segmentation accuracy and convergence speed. Experimental results demonstrate MrisNet achieves high recall and precision rates of 94.8% and 95.2%, respectively, in ship instance segmentation, outperforming various YOLO and other single-stage algorithms. Moreover, MrisNet has a model parameter size of 13.8M and real-time computational cost of 23.5G, demonstrating notable advantages in terms of convolutional efficiency. In conclusion, MrisNet accurately segments ships with different spot features and under diverse environmental conditions in marine radar images. It exhibits outstanding performance, particularly in extreme scenarios and challenging interference conditions, showcasing robustness and applicability.

show abstract

Section: Ship Identification Methods Under Radar and Other Scenariosmentioning

confidence: 99%

MrisNet: Robust Ship Instance Segmentation in Challenging Marine Radar Environments

Ma,

Kang,

Chen

et al. 2023

JMSE

View full text Add to dashboard Cite

show abstract

“…MOTA reflects the number of false detection, missed detection and target ID switching in the process of moving target tracking, which can be expressed as tracking accuracy, the specific expression is as follows ( 10).…”

Section: /12mentioning

confidence: 99%

“…The representative algorithms include Fast R-CNN 8 , Faster R-CNN 9 , etc. For example, Chen Xiaolong et al 10 proposed a target detection algorithm based on Marine Faster R-CNN which has strong generalization ability when the network structure is more complex.In the field of target tracking, Alex et al 11 used the sort algorithm to predict the target motion trajectory using Kalman filter, and combined the Hungarian algorithm to match the front and rear states. The target tracking speed is fast, and the disadvantage is that it is not suitable for occlusion tracking.…”

Section: Introductionmentioning

confidence: 99%

Moving targets intelligent detection and tracking algorithm for that can restrain of local occlusion

YANG,

Hu,

Zhang

et al. 2023

Preprint

View full text Add to dashboard Cite

The optical imaging of the target motion process is susceptible to occlusion, resulting in detection failure and tracking loss. Aiming at the problems of low detection accuracy and poor robustness of existing multi-target tracking algorithms, this paper improves the YOLOv8 detection algorithm and DeepSort tracking algorithm, and constructs a YOLOv8 + DeepSort multi-target detection and tracking joint algorithm. In order to enhance the feature extraction ability of the model, the CBAM module is introduced into the YOLOv8 detection algorithm to retain more feature information. In order to further improve the tracking performance of the DeepSort algorithm, the Hungarian algorithm is selected to achieve the optimal matching of the detection box and the prediction box. The Euclidean distance is used to replace the IOU for the second matching of the failed detection box. The improved YOLOv8 and DeepSort joint algorithm is verified by experiments. The experimental results show that the average accuracy of the detection algorithm based on the improved YOLOv8 reaches 99.5 %, and the comprehensive detection and evaluation data are better than YOLOv4-tiny, YOLOv5, and YOLOv7 algorithms., The moving target is tracked in combined with the improved DeepSort, such that the Multiple Object Tracking Accuracy (MOTA) and Multiple Object Tracking Precision (MOTP) can reach 66.1 % and 84.3 % respectively, and the count of ID conversion is reduced by 16, verifying the effectiveness of the algorithm.

show abstract

“…The accuracy and real-time performance of its detection effect are important indicators to measure machine vision. At present, object detection has been widely used in many fields such as automatic driving, fault detection, smart agriculture, posture detection, radar navigation and so on [2][3][4][5][6] .…”

Section: Introductionmentioning

confidence: 99%

Using the improved YOLOv5 network to detect the real-time and effective position of milk and construct the milk data set

Xia

Wang

Ran

et al. 2023

Preprint

View full text Add to dashboard Cite

In order to improve the detection speed of YOLOv5(You Only Look Once v5) in complex environments and dense target scenarios, a target detection method CN-YOLOv5(Cow Milk-You Only Look Once v5) improved YOLOv5 model is proposed. The traditional YOLOv5 network structure is improved, and the ability of the algorithm to extract features is improved by adding the SE (Squeeze and Excitation) attention module structure, and the accuracy of milk identification is improved. By improving the SPP (Spatial Pyramid Pooling) structure to SPPF (Spatial Pyramid Pooling Fast) structure, the detection speed is accelerated, and the CN-PAN (Cow Nipple Path Aggregation Network) model is proposed based on the PAN (Path Aggregation Network) module. Based on the PAN structure in the traditional YOLOv5 network, the iteration of small target detection is lightweight. Based on YOLOv5s, the milk image dataset CNmodel-YOLOV5s(Cow Milk model-You Only Look Once v5) was created. Experimental results show that the two algorithms can be tested before and after the improvement by using the milk dataset CNmodel-YOLOV5s. The improved algorithm on the test equipment increases the detection speed by up to 13% with almost no impact on accuracy. The improved YOLOV5 algorithm can identify milk targets more quickly, which provides theoretical support for subsequent detection of medium and large targets in complex environments and dense target scenarios.

show abstract

Marine target detection based on Marine-Faster R-CNN for navigation radar plane position indicator images

Cited by 14 publications

References 19 publications

MrisNet: Robust Ship Instance Segmentation in Challenging Marine Radar Environments

MrisNet: Robust Ship Instance Segmentation in Challenging Marine Radar Environments

Moving targets intelligent detection and tracking algorithm for that can restrain of local occlusion

Using the improved YOLOv5 network to detect the real-time and effective position of milk and construct the milk data set

Contact Info

Product

Resources

About