Picking method of slightly rough exposed section based on quadratic fitting algorithm of peak detection and raster
The automatic detection technology of exposed cross-sections is the basis for intelligent control of the suspension height of airborne-coupled ground-penetrating radar antennas.When low-frequency airborne-coupled ground-penetrating radar is used for long-distance exploration in underground coal mines,the exposed cross-sections exhibit characteristics of approximately horizontal straight lines.Currently,it mainly relies on expert manual interpretation,which cannot meet the industrialized requirements of real-time,intelligent,and efficient applications.Therefore,a micro-rough exposed cross-section automatic detection method is proposed based on peak detection algorithm and grid-based quadratic linear fitting method.A window-constrained average method is proposed to effectively filter out direct waves while preserving the detailed response characteristics of the target interface.The time-varying deconvolution method is used to suppress the convolution effect caused by overlapping waveforms to improve waveform quality.A peak detection algorithm based on the idea of local extremes is proposed to automatically extract the peak position information in the A-scan graph of ground-penetrating radar.A grid-based quadratic linear fitting method is proposed to map the peaks to a two-dimensional grid and perform linear fitting.Based on the correlation coefficient and first-order derivative of the fitting curve,the curves are classified,and secondary linear fitting is performed to automatically extract the position labeling and coordinate information of layered target interfaces.Engineering experimental methods are used to verify the applicability of this method.
air coupled ground penetrating radarheight of suspensionbare sectionpeak detectionrasterized linear fitting
NETL
NSTL
万方数据
