Detection And Analysis Of Semivariogram Of Surface Spatial Structure Based On Remote Sensing Image

With the widespread application of multi-spectral sensors,it is increasingly mature to extract surface information by terrain spectral response characteristics technology.However,due to the complexity of the surface conditions and the limitations of spectral response,traditional spectral analysis is flawed in indicating pattern information such as average size and spatial distribution.It has been found that there is a certain relationship between the variogram parameters and the surface spatial structure parameters,However,the current research on this relationship is mostly based on the real image lacking integrity and the series.Due to the short of comprehensive of the variogram detection performance,the current effect of variogram to detect the crown size of remote sensing image is still in doubt.Based on a series of one-dimensional and two-dimensional simulated images of different size distributions and spatial distributions,the variogram values are calculated and the range parameter are extracted.By analyzing the relationship between the range parameter and the size of object in the simulated image,we have discussed the factors influencing the effect of the range parameter to detect object size,and the accuracy of the object size is explored under different spatial structure conditions.Finally,the conclusion of the simulation image analysis is verified by the real images of the South of France on the Mediterranean coast and central Spain.The results show that:(1)When the crown diameter of the forest is less than half of the average distance between the canopy and the variance of the canopy size is small,the range parameter can effectively detect the forest canopy size in the remote sensing image;(2)When the size of the forest crown is greater than half of the average distance between the canopy,the range parameter does not increase with the increase of the canopy size;(3)When the variance of variance of tree crown is larger and the crown diameter is less than half the average distance between canopy,the size of the crown detected by the inflection point of the variogram is smaller than its true value,while the peak of the variogram curve to detect the size of the crown is greater than its real value,and this deviation increases as the canopy size variance increases.