Research On Leaf Area And Standing Tree Measurement Algorithms Based On Image Analysis Embedded In Smart

Forest information collection is the basis of forest resources investigation and forestry visualization, which is a complex and arduous task."Digital forestry" and "precision forestry" put forward higher requirements of forest information collection methods. Basides consuming a large a nomunt of manpower and finacial resources, mannal reading and recording needed in the traditional forest measurement method has a low detree of automation and precision. Many auxiliary measuring tools having appeared in recent years, on the one hand, are expensive and inconveniently-acrried, on the other hand, they might result in severe damaged to the forest in the course of the investigation.At present, mobile phone camera has reached a very high level performance. By using the mobile phone camera as the image acquisition device and the corresponding programs embedded into the mobile phone, real-time measurement data such as plant leaf area, tree height, tree diameter or volume can be measured. The measured data and the GIS information can be updated to the data center by wireless communication and computer network, or can be exchanged with the data center, which provides new methods and approaches to the forest resources investigation.In this thesis, basic researches were made on new methods and approaches based on the image analysis embedded in smart mobile phone. The main work and contributions are as follows:1. A leaf area measurement method based on the similarity color channel is proposed. Beside the rapid speed, the method can achieve a very high accuracy because it overcomes the influence of the shadow on the leaf area, with which leaf area can be segmented by the use of the R, G and color channel similarity, thus the leaf area is calculated according to its pisel number and referenced rectangular pisel number.2. A standing tree image segmentation algorithm based on MRF in complex background is improved. With which the trimap is identified by making full use of the powerful interactive function of the smart mobile phone and its multi-touch technology. As the foreground, background and the unknown region are exactly marked out, and the computation cost is greatly reduced.3. A camera self calibration method based on two images is put forward. Two markers are set on tree trunk before photographed. After two images of the tree are shot in different sites with a certain angle interval, the markers are detected. Then the second image is zoomed according to the Euclidean distance of the two markers on the first image to make the Euclidean distance of two markers in two images equal. Thus, the two shooting locations can be on the circumference of the detected tree center and the intrinsic parameters of camera can be calibrated.4. Some methods for the calculating of the tree height, diameter and volume are presented based on image analysis. The tree height measurement in this thesis is divided into automatic measurement under branch height and arbitrarily high measurement of trunks. Tree diameter measurement method can measure the diameter at breast or at any height of the tree trunk, and the standing tree volume measurement can be made with measuremental method by section.