A Salt Distribution Sensing Method Using Radar Sensor and Neural Networks

The size distribution of the salt particles plays a crucial role in desalination. Knowing the size range of the salt allows us to select appropriate operational conditions for the salt separation process, ultimately leading to improved separation efficiency. Although methods such as using the microscope and laser diffraction offer accurate results for particle size analysis, they often suffer from time-consuming procedures and the need for additional treatment of solid particles. This study introduces a contactless and convenient method to classify the size ranges of salt particles using a radar sensor and Neural Networks(NNs). The sensor captured the reflected radio signals from the samples of three different size ranges. Continuous Wavelet Transform (CWT) was utilized to extract the time-frequency information from the collected data. Deep Convolutional Neural Networks(CNNs) were applied to classify the processed data. Instead of designing and training a neural network from scratch, some pre-trained networks were utilized, which is known as transfer learning. The achieved accuracy reached up to 77.78%, demonstrating the effectiveness of the approach.