The radar shadowing, or masking, is a well known problem in the radar community. This happens whenever multiple targets are located in the shadowing region of other targets, in which they can be detected with lower reliability. In this paper, a novel solution for the radar shadowing problem is proposed. The solution is based on a CNN model that takes as input the spectrograms obtained after a Short Time Fourier Transform (STFT) analysis and classifies the radar output among two classes: Single or Two targets. The model is based on pre-trained MobileNet. The proposed solution achieves a testing accuracy of 88.7% with a standard deviation of 2.39%. The trained model is considered a light model. It only has 1.06 million parameters, and the inference time using a GPU is 1.64ms.
Our custom dataset need to have the following structure: for every class create a folder containing .jpg sample images:
dataset_folder\
class1\
image1.jpg
image2.jpg
class2\
image1.jpg
image2.jpg
image3.jpg
For any help, please contact me: https://www.linkedin.com/in/ammar-mohanna/
- Configure the parameters in config.json
- Train and evaluate the model using
python train.py