Research On Compressed Sensing Method Of W-band Active Millimeter-wave Imaging

In today's world,although most countries and regions are far away from war,terrorist attacks are not far away from everyone.Human security screening in crowded places has become an important way to contain terrorist incidents.Millimeter-wave imaging technology has gradually become an important development and research technology in the field of security inspection due to its many advantages.However,the array of traditional active millimeter wave imaging systems is very dense,the amount of data collection is large,the system complexity and cost are high,so it is of great significance to reduce the number of array elements and the amount of data collection.In this paper,millimeter wave imaging technology based on compressed sensing(CS)is studied and combined with sparse array design.The main research contents of this paper include:This paper introduces the single-frequency millimeter-wave holographic imaging technology,derives the two-dimensional millimeter-wave accurate holographic reconstruction formula,gives the image evaluation index,discusses the sampling interval,sampling range and resolution of the imaging system,and performs simulation verification.Two kinds of millimeter-wave imaging models based on CS are derived,which are CS model based on back propagation(BP)algorithm and CS model based on Fourier transform.Then,the millimeter wave image reconstruction algorithm based on full variational regularization is used for simulation,the imaging differences of different models and the influence of mesh size on the imaging performance of CS model based on BP algorithm are analyzed.In this paper,four different sparse sampling matrices are proposed for two-dimensional millimeter-wave imaging and the imaging performance is simulated and analyzed.In this paper,the system model of wide-band millimeter-wave holographic imaging is given,and the holographic imaging algorithm of wide-band millimeterwave is derived.Based on this,the CS model based on Fourier transform is extended to three dimensions,and the CS model of millimeter-wave three-dimensional imaging is derived.In this paper,three different sparse sampling methods are proposed for three-dimensional millimeter-wave imaging,A three-dimensional model is established and a series of simulations are performed,and the difference in imaging performance of different sampling methods are analyzed.In this paper,three different sparse arrays are studied,that is,cross-receiver array,linear sparse array and coprime array.For the three different sparse arrays,the traditional BP algorithm and the compressed sensing algorithm are used to reconstruct the image in this paper.For coprime arrays,this paper attempts to further reduce the number of antennas in the array with the help of compressed sensing algorithms.At the same time,the quality of the reconstructed image is still better than that of the BP algorithm in the single-transmission and single-reception sampling mode.In summary,this paper applies the compressed sensing algorithm to twodimensional and three-dimensional active millimeter wave imaging system,which greatly reduces the amount of data acquisition.This paper combines the compressed sensing algorithm with the sparse array to improve the imaging quality while greatly reducing the number of antenna elements in the imaging system.