Underwater Sea Cucumber Identification Based on Improved YOLOv5

Zhai, Xianyi; Wei, Honglei; He, Yan; Shang, Yetong; Liu, Chenghao

doi:10.3390/app12189105

Cited by 21 publications

(12 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many researchers choose to combine attention mechanisms with neural networks to improve the detection accuracy of underwater targets [34,35]. Sun et al [36] attempted to use the Swin Transformer to design a new network as the backbone for underwater target detection, achieving performance similar to the cascade R-CNN with the ResNeXt101 backbone.…”

Section: Attention Mechanismmentioning

confidence: 99%

DyFish-DETR: Underwater Fish Image Recognition Based on Detection Transformer

Wang,

Ruan,

Chen

2024

JMSE

View full text Add to dashboard Cite

Due to the complexity of underwater environments and the lack of training samples, the application of target detection algorithms to the underwater environment has yet to provide satisfactory results. It is crucial to design specialized underwater target recognition algorithms for different underwater tasks. In order to achieve this goal, we created a dataset of freshwater fish captured from multiple angles and lighting conditions, aiming to improve underwater target detection of freshwater fish in natural environments. We propose a method suitable for underwater target detection, called DyFish-DETR (Dynamic Fish Detection with Transformers). In DyFish-DETR, we propose a DyFishNet (Dynamic Fish Net) to better extract fish body texture features. A Slim Hybrid Encoder is designed to fuse fish body feature information. The results of ablation experiments show that DyFishNet can effectively improve the mean Average Precision (mAP) of model detection. The Slim Hybrid Encoder can effectively improve Frame Per Second (FPS). Both DyFishNet and the Slim Hybrid Encoder can reduce model parameters and Floating Point Operations (FLOPs). In our proposed freshwater fish dataset, DyFish-DETR achieved a mAP of 96.6%. The benchmarking experimental results show that the Average Precision (AP) and Average Recall (AR) of DyFish-DETR are higher than several state-of-the-art methods. Additionally, DyFish-DETR, respectively, achieved 99%, 98.8%, and 83.2% mAP in other underwater datasets.

show abstract

Section: Attention Mechanismmentioning

confidence: 99%

DyFish-DETR: Underwater Fish Image Recognition Based on Detection Transformer

Wang,

Ruan,

Chen

2024

JMSE

View full text Add to dashboard Cite

show abstract

“…Firstly, according to the Stokes vector method, the total intensity image I of 0°, 45°, 90° and 135° polarization images is obtained using Equation (1).…”

Section: Principle Of Polarization Imagingmentioning

confidence: 99%

“…There are problems such as image brightness attenuation and contrast reduction. The problems in these images will bring some difficulties to underwater environment monitoring, military reconnaissance and other target detection activities [1,2]. Polarization imaging is a new optical imaging technology.…”

Section: Introductionmentioning

confidence: 99%

Polarization Imaging Method for Underwater Low-Visibility Metal Target Using Focus Dividing Plane

Xue

Zhao

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

Aiming at the problems of brightness attenuation and contrast reduction in the target image caused by underwater low-visibility environments, a metal target detection method based on split-focus plane polarization imaging is proposed. Firstly, a hybrid enhancement method is proposed to clarify the degraded polarization image. In this study, the GrayWorld method is improved to compensate the attenuation difference of the total light intensity of the polarization image. Variational contrast and saturation enhancement algorithms are used to reduce the underwater scattering effect; secondly, a split-focus plane polarization imaging system is built to complete the control of camera parameters, polarization image acquisition and information processing. Under different underwater low-visibility conditions, polarization imaging of targets with different materials can be realized; finally, an image quality evaluation system is constructed to compare the light intensity and degree of polarization images that are collected by the focal plane polarization imaging system. The polarization characteristics of metal and nonmetal target plates are analyzed. The results show that under the condition of low visibility, the obtained polarization image contrast of the metal target is relatively high, and its EME, information entropy and average gradient are increased by 183.82%, 53.46% and 586.22% on average relative to the image of light intensity. In an underwater low-visibility scene, the method of focal plane polarization imaging proposed in this paper can reduce the difficulty of metal target detection.

show abstract

“…The backbone is the basic feature extraction layer, which is responsible for extracting the feature information of the image; the Neck is the feature fusion layer, which is responsible for fusing the image information of different scales to obtain better detection results; the Head is the output layer, which is responsible for outputting the detected target information. The schematic diagram is shown in figure 6 [18].…”

Section: Yolov5 Deep Learning Networkmentioning

confidence: 99%

Chip detection algorithm based on lightweight E-YOLOv5 convolutional neural network

Zhai

Meng²,

Wei³

2023

Eng. Res. Express

Self Cite

View full text Add to dashboard Cite

To solve the chip location recognition problem, this paper proposes a lightweight E-YOLOv5 based chip detection algorithm based on the You Only Look Once version 5 (YOLOv5s) algorithm. For the problem of the difficult distinction between chip detection points and light spots, a simulated exposure algorithm is used to process part of the training set images to enhance model robustness; the existing model network is complex, and EfficientNet, a lightweight feature extraction network, is introduced to reduce the model size; for the problem of imprecise model recognition due to small detection points, Selective Kernel Neural Network (SKNet) module is introduced into EfficientNet is introduced to enhance the feature extraction ability of the model and improve the training efficiency, and Efficient Intersection over Union Loss (EIoU_Loss) is used as the loss function to reduce the false recognition rate. Experiments show that the algorithm in this paper improves by 3.85% and 3.92% in precision, recall rate, 28.89% in loss value, nearly 20% in model size and training time, and 46.67% in image processing speed on CPU compared with YOLOv5s. The experimental results show that the proposed algorithm outperforms other algorithms and is able to distinguish and identify chip locations precisely and stably.

show abstract

Underwater Sea Cucumber Identification Based on Improved YOLOv5

Cited by 21 publications

References 13 publications

DyFish-DETR: Underwater Fish Image Recognition Based on Detection Transformer

DyFish-DETR: Underwater Fish Image Recognition Based on Detection Transformer

Polarization Imaging Method for Underwater Low-Visibility Metal Target Using Focus Dividing Plane

Chip detection algorithm based on lightweight E-YOLOv5 convolutional neural network

Contact Info

Product

Resources

About