SNR‐dependent drone classification using convolutional neural networks
Abstract Radar sensing offers a method of achieving 24‐h all‐weather drone surveillance, but in order to be maximally effective, systems need to be able to discriminate between birds and drones. This work examines drone‐bird classification performance as a function of signal to noise ratio (SNR). Cl...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2022-01-01
|
| Series: | IET Radar, Sonar & Navigation |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/rsn2.12161 |
| _version_ | 1811273496341250048 |
|---|---|
| author | Holly Dale Chris Baker Michail Antoniou Mohammed Jahangir George Atkinson Stephen Harman |
| author_facet | Holly Dale Chris Baker Michail Antoniou Mohammed Jahangir George Atkinson Stephen Harman |
| author_sort | Holly Dale |
| collection | DOAJ |
| description | Abstract Radar sensing offers a method of achieving 24‐h all‐weather drone surveillance, but in order to be maximally effective, systems need to be able to discriminate between birds and drones. This work examines drone‐bird classification performance as a function of signal to noise ratio (SNR). Classification at low SNR values is necessary in order to classify drones with a small radar cross‐section (RCS), as well as to facilitate reliable classification at longer ranges. To investigate the relationship between classification performance and SNR, Gaussian noise is added to an experimentally obtained dataset of radar spectrograms. Classification is performed by convolutional neural networks (CNNs). It is shown that for the data available classification accuracy drops with falling SNR, as might be expected for any given CNN. The degree to which performance degrades with reduced SNR is presented. It is further shown that simpler network architectures are more robust to noise. Finally, it is demonstrated that data augmentation can be used as a means of enhancing classification accuracy at lower SNR values. Bayesian optimisation is used to find the optimal augmentation hyperparameters and overall, classification accuracies of 92% are achieved at low SNR. |
| first_indexed | 2024-04-12T23:00:11Z |
| format | Article |
| id | doaj.art-76d0d3c25ec14a449baeb4b454cdcd60 |
| institution | Directory Open Access Journal |
| issn | 1751-8784 1751-8792 |
| language | English |
| last_indexed | 2024-04-12T23:00:11Z |
| publishDate | 2022-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Radar, Sonar & Navigation |
| spelling | doaj.art-76d0d3c25ec14a449baeb4b454cdcd602022-12-22T03:13:04ZengWileyIET Radar, Sonar & Navigation1751-87841751-87922022-01-01161223310.1049/rsn2.12161SNR‐dependent drone classification using convolutional neural networksHolly Dale0Chris Baker1Michail Antoniou2Mohammed Jahangir3George Atkinson4Stephen Harman5Microwave Integrated Systems Laboratory University of Birmingham Birmingham UKMicrowave Integrated Systems Laboratory University of Birmingham Birmingham UKMicrowave Integrated Systems Laboratory University of Birmingham Birmingham UKMicrowave Integrated Systems Laboratory University of Birmingham Birmingham UKMicrowave Integrated Systems Laboratory University of Birmingham Birmingham UKAveillant Cambridge UKAbstract Radar sensing offers a method of achieving 24‐h all‐weather drone surveillance, but in order to be maximally effective, systems need to be able to discriminate between birds and drones. This work examines drone‐bird classification performance as a function of signal to noise ratio (SNR). Classification at low SNR values is necessary in order to classify drones with a small radar cross‐section (RCS), as well as to facilitate reliable classification at longer ranges. To investigate the relationship between classification performance and SNR, Gaussian noise is added to an experimentally obtained dataset of radar spectrograms. Classification is performed by convolutional neural networks (CNNs). It is shown that for the data available classification accuracy drops with falling SNR, as might be expected for any given CNN. The degree to which performance degrades with reduced SNR is presented. It is further shown that simpler network architectures are more robust to noise. Finally, it is demonstrated that data augmentation can be used as a means of enhancing classification accuracy at lower SNR values. Bayesian optimisation is used to find the optimal augmentation hyperparameters and overall, classification accuracies of 92% are achieved at low SNR.https://doi.org/10.1049/rsn2.12161optimisationsignal classificationradar signal processingradar cross‐sectionsBayes methodsGaussian noise |
| spellingShingle | Holly Dale Chris Baker Michail Antoniou Mohammed Jahangir George Atkinson Stephen Harman SNR‐dependent drone classification using convolutional neural networks IET Radar, Sonar & Navigation optimisation signal classification radar signal processing radar cross‐sections Bayes methods Gaussian noise |
| title | SNR‐dependent drone classification using convolutional neural networks |
| title_full | SNR‐dependent drone classification using convolutional neural networks |
| title_fullStr | SNR‐dependent drone classification using convolutional neural networks |
| title_full_unstemmed | SNR‐dependent drone classification using convolutional neural networks |
| title_short | SNR‐dependent drone classification using convolutional neural networks |
| title_sort | snr dependent drone classification using convolutional neural networks |
| topic | optimisation signal classification radar signal processing radar cross‐sections Bayes methods Gaussian noise |
| url | https://doi.org/10.1049/rsn2.12161 |
| work_keys_str_mv | AT hollydale snrdependentdroneclassificationusingconvolutionalneuralnetworks AT chrisbaker snrdependentdroneclassificationusingconvolutionalneuralnetworks AT michailantoniou snrdependentdroneclassificationusingconvolutionalneuralnetworks AT mohammedjahangir snrdependentdroneclassificationusingconvolutionalneuralnetworks AT georgeatkinson snrdependentdroneclassificationusingconvolutionalneuralnetworks AT stephenharman snrdependentdroneclassificationusingconvolutionalneuralnetworks |