Detection of Small Floating Target on Sea Surface Based on Gramian Angular Field and Improved EfficientNet
In order to exploit the advantages of CNN models in the detection of small floating targets on the sea surface, this paper proposes a new framework for encoding radar echo Doppler spectral sequences into images and explores two different ways of encoding time series: Gramian Angular Summation Field...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-09-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/17/4364 |
| _version_ | 1797493315524886528 |
|---|---|
| author | Caiping Xi Renqiao Liu |
| author_facet | Caiping Xi Renqiao Liu |
| author_sort | Caiping Xi |
| collection | DOAJ |
| description | In order to exploit the advantages of CNN models in the detection of small floating targets on the sea surface, this paper proposes a new framework for encoding radar echo Doppler spectral sequences into images and explores two different ways of encoding time series: Gramian Angular Summation Field (GASF) and Gramian Angular Difference Field (GADF). To emphasize the importance of the location of texture information in the GAF-encoded map, this paper introduces the coordinate attention (CA) mechanism into the mobile inverted bottleneck convolution (MBConv) structure in EfficientNet and optimizes the model convergence by the adaptive AdamW optimization algorithm. Finally, the improved EfficientNet model is used to train and test on the constructed GADF and GASF datasets, respectively. The experimental results demonstrate the effectiveness of the proposed algorithm. The recognition accuracy of the improved EfficientNet model reaches 96.13% and 96.28% on the GADF and GASF datasets, respectively, which is 1.74% and 2.06% higher than that that of the pre-improved network model. The number of parameters of the improved EfficientNet model is 5.38 M, which is 0.09 M higher than that of the pre-improved network model. Compared with the classical image classification algorithm, the proposed algorithm achieves higher accuracy and maintains lighter computation. |
| first_indexed | 2024-03-10T01:18:16Z |
| format | Article |
| id | doaj.art-988bd5dfe9d14fe38a5e5f619ff53591 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T01:18:16Z |
| publishDate | 2022-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-988bd5dfe9d14fe38a5e5f619ff535912023-11-23T14:05:23ZengMDPI AGRemote Sensing2072-42922022-09-011417436410.3390/rs14174364Detection of Small Floating Target on Sea Surface Based on Gramian Angular Field and Improved EfficientNetCaiping Xi0Renqiao Liu1College of Automation, Jiangsu University of Science and Technology, Zhenjiang 212100, ChinaOcean College, Jiangsu University of Science and Technology, Zhenjiang 212100, ChinaIn order to exploit the advantages of CNN models in the detection of small floating targets on the sea surface, this paper proposes a new framework for encoding radar echo Doppler spectral sequences into images and explores two different ways of encoding time series: Gramian Angular Summation Field (GASF) and Gramian Angular Difference Field (GADF). To emphasize the importance of the location of texture information in the GAF-encoded map, this paper introduces the coordinate attention (CA) mechanism into the mobile inverted bottleneck convolution (MBConv) structure in EfficientNet and optimizes the model convergence by the adaptive AdamW optimization algorithm. Finally, the improved EfficientNet model is used to train and test on the constructed GADF and GASF datasets, respectively. The experimental results demonstrate the effectiveness of the proposed algorithm. The recognition accuracy of the improved EfficientNet model reaches 96.13% and 96.28% on the GADF and GASF datasets, respectively, which is 1.74% and 2.06% higher than that that of the pre-improved network model. The number of parameters of the improved EfficientNet model is 5.38 M, which is 0.09 M higher than that of the pre-improved network model. Compared with the classical image classification algorithm, the proposed algorithm achieves higher accuracy and maintains lighter computation.https://www.mdpi.com/2072-4292/14/17/4364radar target detectionsea clutterimage classificationdeep learning |
| spellingShingle | Caiping Xi Renqiao Liu Detection of Small Floating Target on Sea Surface Based on Gramian Angular Field and Improved EfficientNet Remote Sensing radar target detection sea clutter image classification deep learning |
| title | Detection of Small Floating Target on Sea Surface Based on Gramian Angular Field and Improved EfficientNet |
| title_full | Detection of Small Floating Target on Sea Surface Based on Gramian Angular Field and Improved EfficientNet |
| title_fullStr | Detection of Small Floating Target on Sea Surface Based on Gramian Angular Field and Improved EfficientNet |
| title_full_unstemmed | Detection of Small Floating Target on Sea Surface Based on Gramian Angular Field and Improved EfficientNet |
| title_short | Detection of Small Floating Target on Sea Surface Based on Gramian Angular Field and Improved EfficientNet |
| title_sort | detection of small floating target on sea surface based on gramian angular field and improved efficientnet |
| topic | radar target detection sea clutter image classification deep learning |
| url | https://www.mdpi.com/2072-4292/14/17/4364 |
| work_keys_str_mv | AT caipingxi detectionofsmallfloatingtargetonseasurfacebasedongramianangularfieldandimprovedefficientnet AT renqiaoliu detectionofsmallfloatingtargetonseasurfacebasedongramianangularfieldandimprovedefficientnet |