Forest Canopy Porosity Extraction Method

Vegetation completed change of various of substances by carrying on Photosynthesis, Respiration, Transpiration and so on.The leaf blade of green foliage mainly undertake change.The density,distribution of leaf directly influence cycle of C,N,O and water.The quantitative descripition of vegetation leaf reflects the local ecosystem.The quantitative analysis and research of vegetation leaf mainly includes the reasearch of leaf area index.The gap fraction is one of inportant parameters when LAI is calculated.Fisheye camera is used for obtaining the image of gap fraction.As the image will distort,the fisheye camera need be demarcated befor studying image.Through the internal parameters of camera and utilizing the character of the distortion factor,the image is corrected.This research put forward the promoted of Susan operator through using the method of 4 corners of the neighborhood template.Using this template to scan image.The number of times of gray level jumping is 0 in the flat areas.The number of times is 2 at the edge point.If the point is corner,the number of times will be grater or equal to 2.1f the point is nois points(also called isolated point),The number of times will be O.As the number of times is grater or equal to 2,Susan operator will be used for detecting image.With regard to segmentate and extract of forest canopy gap fraction,this research put forward 2 segmentation algorithms.The first one is promoted OTSU algorithm.The image pre-splited is called grayscale image after being extracted blue component under the RGB image.According to the difference between zenith angle and azimuth angle,concentric ring function and sector model,the image is dived into 18 concentric rings and fan-shaped image.The gray value of pixels of image pre-splited is extraced and determined the position of frame pixels through Matlab.Then the grayscale value of this pixel is excluded.The information of remaining pixels is grayscale of target area.The threshold of OTSU of this set of data can be acquired. The second is gray threshold algorithm based on standard template, first taking a photograph in the broad-leaved forest forest near the open sky, taking 3 photograghs of broad-leaved forest, taking a photograph of the empty sky near the coniferous forest,3 photograghs of coniferous f orest, graying the two pieces of open wide sky respectively and use the average grayscale as a s tandard template threshold, then doing difference comparison between all points of tested imag e and standard template threshold, the final segmentation threshold is the average value of all t he difference. The image of forest canopy porosity can be obtained by extracting the points whi ch are larger than the average value.