Evaluation of the Distance to Thermal Objects by Passive MWIR Sensors: A Multidimensional Regression Approach

An approach to determining the distance to thermal objects on the basis of data obtained from a mid-wave infrared camera is used. The radiative transfer equation is analyzed, and the statistical stability of the target thermal distribution is demonstrated depending on the distance to the object and weather conditions. Based on the set of identified object features, a multivariate linear regression model is proposed that uses a trained artificial neural network and can predict the distance to the object. Known experimental data have been used to model this approach, showing the advantage of the estimated distance (less than $$\Delta_{D}/D<0.04$$ ) in comparison with the direct method of determining the distance ( $$\Delta_{D}/D>0.82$$ ). Lines for further research are given, and the need to improve the feature space aimed at increasing the accuracy of this approach is substantiated.