Research And Implementation Of Forest Canopy Hemisphere Image Segmentation Methods

Forest canopy structure parameters is important to the evaluation the ecological environment, carbon sink measurement and studies of vegetation growth, forest canopy hemisphere image segmentation, is one of the important ways to obtain forest canopy structure parameters, the choice of image segmentation method and obtain the image partly due to the reflection of vegetation seriously affect the accuracy of image segmentation and expensive commercial canopy hemisphere image acquisition device, aiming at these problems, this paper research content is as follows:(1) The vegetation reflective phenomenon, cause local too bright, seriously affect the accuracy of image segmentation, and put forward to eliminate the image of local light phenomenon, and analyze the results of image segmentation of reflection and not reflection.(2) Using the the traditional method of edge detection method,threshold method, region growing method to segment canopy hemisphere image, and analyzing the experimental results, combining the statistical methods of density, maximizing minimum distance method and k-means clustering method, improving the k-means clustering algorithm, using the improved k-means clustering algorithm to segment the canopy hemisphere images, analyzing the segmentation result and the length of time of program running.(3) In order to improve the segmentation accuracy of the canopy hemisphere image, combining the Particle swarm optimization and k-harmonic clustering algorithm, improving the k-harmonic clustering algorithm, using the improved k-harmonic clustering algorithm to segment the color canopy hemisphere images, analyzing of the experimental results, comparing the effects of segmentation way to the the length of time of program running.(4) Designing canopy hemisphere image acquisition system hardware structure, using servo drive the mechanical structure, control the camera position and angle, drawing the circuit principle diagram; Design the mechanical movement control and image acquisition upper computer, we can use the upper computer control movement and take photos, made a image processing interface, used for image preprocessing and image segmentation, and realized the image processing method in this paper.