(If you are benefited from this dataset, please cite us.) It can be downloaded from and extract by (MFGJ_Dataset.zip):
The MFGJ dataset is divided into a training set and a test set with a JSR in the range of [15db, 50db],where there are a total of 46,800 spectrogram images in the training set and 21,600 spectrogram images in the test set, for a total of 68,400 spectrogram images.
Train\Test | 00Clean | 01CWI | 02DME | 03Hooked- Sawtooth |
04Linear | 05Triangular | 06Triangular- Wave |
07Sawtooth | 08Tick | 09DSSS | 10NBFMJ | 11BLGNI |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Train | 3600 | 3600 | 3600 | 3600 | 3600 | 7200 | 3600 | 3600 | 3600 | 3600 | 3600 | 3600 |
Test | 1800 | 1800 | 1800 | 1800 | 1800 | 1800 | 1800 | 1800 | 1800 | 1800 | 1800 | 1800 |
Annotations are embedded in file name. A sample image name is "image00001_01CWI_jsr15", Each name can be splited into THREE fields. Those fields are explained as follows.
- image_index:The index number of the image under the corresponding category
- type: Classes of jamming in GNSS
- JSR: JSR of the corresponding image
<type = ["00Clean", "01CWI", "02DME", "03HookedSawtooth", "04Linear", "05Triangular", "06Triangular-<br>Wave", "07Sawtooth", "08Tick", "09DSSS", "10NBFMJ", "11BLGNI"]>
<JSR = [15, 50]>
We publish the code for generating MFGJ datasets in the MFGJdata_Generate folder, so that you can generate a customized number of MFGJ datasets according to your needs.