| The canopy is an important part of the forest,timber forest growth,biomass,wood can be estimated based on canopy.So,the right evaluation of forest canopy height in effective methods and advanced technologies is the important foundation of understanding the pattern,the process of the whole forest ecosystem and its operation mechanism,and also can provide reliable basic data and reference in the subsequent use of forests to improve the environment.The emergence and development of the large-footprint LiDAR technology,to some extent,have achieved quick,without damage estimation of the forest canopy height in the big time and spatial scales,and also have greatly improved the efficiency and accuracy of the forest canopy height estimation.However the light spot size of the spaceborne large-footprint LiDAR is large and susceptible to environment.The topographic slope covered in the light spot is big degree and the terrain isn't flat which interfere with the echo information of the large-footprint LiDAR and cause the estimation deviation of the forest canopy height.So,it is crucial to probe into new denoise method and recede influence of slope to estimation of the forest canopy height.This paper selects the area administered by Wang Qing forestry bureau in Jilin province as the research area and has probed into the ICESat-GLAS data de-noising with hamming window function and gauss window function.This paper also has simulated ICESat-GLAS.echo waveform based on the theory of interaction between laser beam and the reflector and then inverses the model of the maximum forest canopy height based on the denoising waveform and analog waveform.The main research contents and conclusions are as follows:(1)Denoising ICESat-GLAS echo data based on different window function.Because the ICESat-GLAS original waveform contains a lot of noise,this paper put forward to denoising ICESat-GLAS echo waveform using both hamming window function and gauss window function.The results show that the ICESat-GLAS echo waveform is smoother compared with the original waveform figure after denoised by both hamming window function and gauss window function.while the waveform denoised by hamming window function is smoother with the same GLAS waveform data.The denoise method with both hamming window function and gauss window function can reduce the noise effectively,increase the SNR and reduce the mean square error.The denoise method with hamming window function has higher SNR value and lower RMSE value compared with the denoise method with gauss window function in the same slope group.The SNR value is higher and the RMSE value is lower after denoising with the slope increasing and the more rough of the waveform,the better effect of denoising.(2)Simulating ICESat-GLAS echo waveform simulation using the theory of laser radar beam and the interaction of the laser beam and the reflector.This paper simulate ICESat-GLAS echo waveform in different slop with the laser radar transmitted pulse theory and laser pulse with reflector interaction theory.The result shows that the basic configuration of simulated echo waveform figure coincides to the corresponding actual echo waveform figure and the highest Pearson's correlation coefficient is 0.958 while the lowest is 0.886.The consistent results between simulation waveforms and actual waveforms are slightly different but the correlation coefficient isn't quite different which indicates the simulation is effective to some extent.(3)The inversion of the maximum forest canopy height model.When the terrain slope in the model are 0?5°,0?10°,0?15°,0?20°,0?25°,0?30°,the adjusted coefficient of determination R2 are 0.911,0.878,0.769,0.706,0.652,0.563,and the correlation coefficient r is 0.909,0.877,0.831,0.804,0.742,with the inversion of the maximum forest canopy height based on the denoising echo waveform data and topographic index.When using the inversion of the maximum forest canopy height based on the denoising echo waveform data and the simulated echo waveform data in the same slope range,the adjusted coefficient of determination R2 are 0.913,0.884,0.783,0.723,0.680,0.602,and the correlation coefficient r are 0.942,0.915,0.886,0.842,0.822,0.769.(4)The contrasting of two maximum forest canopy height models.Under the same terrain slope,the result basing on the denoising echo waveform data and the simulated echo waveform data is better than the other.When the terrain slope in the model is 0?5°,0?10°,0?15°,0?20°,0?25°,0?30°,the differences of R2 are 0.002?0.006?0.014?0.017?0.028?0.039,the differences of r is 0.001?0.006?0.009?0.011?0.018?0.027.From the comparison results,we know that the results of the maxinum canopy height extimation basing on the denoising echo waveform data and the simulated echo waveform are all better than the other. |
PDF Full Text Request