Local detrended fluctuation analysis for spectral red-edge parameters extraction

Nowadays, spectrum technology makes it possible to nondestructively monitor crop and provide real-time diagnostic advice. Spectral red-edge parameters are important detection target for the growth of rapeseed. In this work, we propose a new method to acquire the rapeseed’s spectral red-edge parameters based on local detrended fluctuation analysis (L-DFA). In practice, by using the L-DFA, local Hurst exponent of the spectrum is calculated firstly. And then, six traditional red-edge parameter extraction methods, namely maximum first derivative method, adjusted linear extrapolation method, linear four-point interpolation method, inverted Gaussian method, Lagrangian interpolation method, and polynomial fitting interpolation, are employed to act on the estimated local Hurst exponent, and thus four red-edge parameters, i.e., red-edge position, red-edge amplitude, and left and right red-edge area, are obtained. In our experiments, by using the above four local Hurst exponent-based red-edge parameters as argument, a prediction model is developed to forecast the SPAD values (soil and plant analyzer development, often used as a parameter to indicate the chlorophyll content) for the five rapeseed’s growth periods based on random decision forest method. The results show that the local Hurst exponent-based red-edge feature can produce better model effect than the original spectrum-based one for the all phenological periods. Significance test for the four kinds of plant patterns shows that there is huge difference of the proposed four-dimensional red-edge features between the transplant and direct plant in the whole process of rapeseed’s growth. The differences between the other three groups of plant factors (two planting densities, two fertilizer types, and two weed controls) are not significant in certain specific growth periods of rapeseed. These findings provide a basis for studying difference of rapeseed yield and oil differences between different planting patterns.