Research On Remote Detection And Inversion Method Of Underground Strip Targets

In recent years,with the increasing demand for land,the means and technology of using underground space have also been progressively improved.The use of underground space has become an important way of urbanization.Economically,the detection of underground targets will assist in the planning,construction,and utilization of underground space.In the military,remote sensing of the enemy’s underground military targets can effectively enhance China’s national defense strength.At present,the detection methods for underground targets have been expanded to include multiple fields.Among them,infrared remote sensing detection methods have obvious advantages in theoretical basis,technical maturity,detection flexibility,and detection cost.Therefore,in this paper,infrared is used as the main method to study the detection and inversion methods of underground strip targets.At the same time,the use of hyperspectral and visible light to conduct preliminary detection of underground facility vents.First of all,based on heat transfer related theory,this paper analyzes and proposes a superposition model of infrared information field disturbance of underground targets and background,and verifies its detectability through mathematical and physical model analysis.Secondly,according to the two detection interference factors of solar radiation and surface vegetation,the solar radiation model and the surface vegetation heat transfer model are obtained respectively to provide background information for filtering interference.According to the principle of disturbance superposition,the satellite infrared remote sensing data and the simulated mountain infrared data are used to obtain the disturbance distribution of the underground strip target.According to the parameters in the relevant specifications of the tunnel construction,the maximum possible offset of the underground strip target is obtained.For the criterion,a filtering enhancement method based on the distance criterion is proposed to filter the target disturbance,and the Gaussian curvature filtering method is introduced to filter out the noise in the enhanced disturbance distribution to obtain a concentrated distribution image of the target disturbance.A tunnel detection method based on position correlation is proposed.The entrance and exit of the tunnel are used as endpoints,and the location and direction of the underground strip target are determined section by section at a certain directional interval.Finally,the detection result is obtained by least square fitting,and the error of the detection result is less than 5%.Thirdly,according to the difference between the material properties of the vent and the surrounding environment,the hyperspectral data differences between the two are analyzed,and the surface is classified to obtain the suspected vent area using hyperspectral.Further use high-resolution visible light data to classify and manually interpret the suspected area to obtain the actual position of the vent.Finally,the optimization theory is used,taking the detected underground strip target position as a reference,the square of the target disturbance as the target function,and iteratively retrieving the target’s burial depth information through the conjugate gradient method,and the error of burial depth is less than 1m.