Soil water content influences soil physical, chemical, and biological processes in soil. Time domain reflectometry (TDR) is a well-established electromagnetic technique used to measure in situ soil water contents nondestructively and continuously. Tangent line algorithms are widely used to determine the reflection positions of TDR signals. A TDR waveform can be influenced by short probes, dry soil and soil salinity, and the accuracy of tangent line methods is limited by such factors. Curvature analysis provides an alternative way for determining reflection positions. The physical assumption for curvature analysis is that the electromagnetic wave passing through the boundary of materials with different dielectric constants will lead to a non-smooth change in TDR waveforms. The non-smooth changes will be used to determine the reflection positions.

 In this study, a curvature analysis algorithm based on Bounded Mean Oscillation (BMO) quantity of TDR waveforms was developed as a new method for determining the reflection positions. A pretreatment based on the Perona-Malik model was used to smooth the TDR waveforms and to improve the effectiveness of curvature analysis. Numerical simulations and experiments with TDR waveforms were used to represent the advantages and disadvantages of the method. The results were used to show that the local maxima in the curvature of waveforms corresponded to the reflection positions. After pretreatment, the local maxima of the curvature were determined easily. The reflection positions determined by curvature analysis were reasonable.