Techniques To Detect Underground Rhizomes

Based on requirements of an enterprise project, the detection technology of underground plant root is studied in this paper. The main tasks include:(1) nondestructively to detect object underground;(2) to identify the underground plant root;(3) to determine the size of plant root. In this paper, ground-penetrating radar is used to nondestructively detect underground plant root, According to the obtained echo data of ground-penetrating radar, the main point in this paper is placed on echo signal processing and underground plant root identification.Firstly, according to the requirements of plant root detection, this paper analyzes electromagnetic wave propagation characteristics in soil and the resolution of ground-penetrating radar and so on; At the same time, electromagnetic characteristics of soil and plant root also are analyzed in order to simulate the detection process of underground object with ground-penetrating radar simulation software.Secondly, the ground penetrating radar signal process is studied. These contents mainly include the Gaussian noise reduction, underground object detection, removal of the direct wave, underground object size estimation, ground-penetrating radar ground echo arrival time estimation. After signal processing, On the one hand, direct wave and noise are suppressed, thereby the echo information of underground object enhanced significantly; On the other hand, the depth and size information of underground object are obtained.Again, underground section image is reconstructed with the two-dimensional echo data of ground penetrating radar.Two different methods of image reconstruction are used in this paper, one is based on characteristics of objects hyperbolic, and another is based on stolt interpolation. After image reconstruction, the azimuth resolution of GPR has been significantly improved.Finally, the identification of underground plant root is studied. Ground-penetrating radar simulation software is used first to obtain the echo data of plant root, stone and other objects, then feature vectors are extracted from these echoes, and support vector machine is adopted to classify these feature vectors. Results show that the recognition accuracy rate of plant root is above90%. Thereby, the identification method described here can be used in the research of identification of different underground objects.