The Influence of Target Orientation on the Underground Targets Classification

Underground target detection and classification have been widely used in various fields. With the development of industrialization, The require for accurate classification of underground targets in various fields is becoming more and more urgent. The orientation of the underground targets are unknown and has an influence on the classification result, Therefore, it is necessary to study the influence of orientation on classification. In this paper, the sensitivity analysis of the underground object model (approximate forward model) was performed to obtain the influence of the object orientation on the model’s output. The influence of the target orientation on the classification results is analyzed by sensitivity analysis results. Next, eight machine learning algorithms are utilized to classify the orientation-known data set and the orientation-unknown data set. Classification results were evaluated using classification accuracy to demonstrate the results of sensitivity analysis. According to the results of the sensitivity analysis, we found that the influence of the orientation parameters on the model’s output varies with the object properties (shape and radius). Therefore, we infer that it is not feasible to classify targets when their orientations are unknown. In the classification results, the classification accuracy of various classification algorithms in the orientation-unknown situation is much lower than that in the orientation-known situation.