Research On Vegetation Information Extraction Technology Based On Digital Image

The vegetation index is an effective measure of the growth status of plants and crops in the same area,such as forests,mountains and farmland,and can reflect the reflection of near-infrared wavelengths and the difference between visible light wavelengths and soil.In this thesis,the image segmentation technology based on wavelet transform is applied to the hemispheric photographic vegetation parameter measurement method to solve the difficult problem of canopy information extraction due to the complex grayscale information of vegetation images.The Leaf Area Index and Diffuse Non-interceptance of vegetation such as Chinese fir,swamp pine and rice were extracted by the vegetation parameter extraction algorithm based on hemispheric photography method and iterative optimization method proposed in this thesis.The measurement results show that the measurement results of the algorithm in this thesis have a good correlation with the reference value.In this thesis,four aspects of vegetation image processing technology,vegetation image segmentation,vegetation parameter extraction based on gap ratio,and algorithm validity verification are studied.The specific research contents are as follows:(1)The image processing methods used in hemispheric photography,namely effective area extraction,lens correction,vegetation image preprocessing,and vegetation image segmentation,are compared and analyzed.The validity of image segmentation technology based on wavelet transform in vegetation image segmentation is affirmed.(2)The advantages and disadvantages of traditional wavelet image segmentation technology applied to vegetation image segmentation are emphatically studied,and the grayscale and frequency characteristics of leaf area,sky background area,and trunk area of vegetation image are analyzed.Based on the canopy characteristics of forest vegetation,four improved schemes for traditional wavelet image segmentation technology are gradually proposed.Among them,the iterative optimization method segmentation scheme significantly improves the vegetation image segmentation effect of traditional wavelet image segmentation technology,and effectively reduces the misjudgment rate of different regions of vegetation images.(3)The inversion technology of vegetation parameters based on canopy gap ratio is studied.The principles of extracting canopy gap ratio from vegetation images and inverting vegetation parameters from canopy gap ratio are deduced in detail.Combined with the iterative optimization method image segmentation scheme,a new set of leaf area index and unintercepted scattering extraction scheme is introduced.(4)The algorithm in this thesis is used to extract the vegetation parameters of the forest vegetation in Qianyan Prefecture and the rice vegetation in Luancheng Station,and the obtained results are analyzed with the measurement results of the LAI-2200 C canopy analyzer.The data analysis results show that for the forest canopy,the Pearson correlation coefficient between the LAI extracted by the algorithm in this thesis and the measured value of the control platform reaches 0.95.The Pearson correlation coefficient of DIFN extracted by the algorithm reached 0.92.The Pearson correlation coefficient of LAI extracted for sparse rice vegetation reached 0.94.The Pearson correlation coefficient of DIFN extracted from rice vegetation reached 0.96.