Application Of Marked Point Process In Crown Extraction From LIDAR Point Cloud

The airborne Light Detection and Ranging(LIDAR) is a new fast measurement system, which makes use of laser rangefinder, IMU, GPS and some other many surveying and mapping technology. The LIDAR technology has some advantages, the speed of data acquisition, short data processing cycle, high-precision, high-density, etc. It brings a band-new technological means for direct, fast three-dimensional data acquisition and can be used in many areas such as digital city construction, topography, vegetation monitoring and geological disaster monitoring.In the digital city construction, the extraction of urban surface information is particularly important. The airborne Light Detection and Ranging system can acquire the information. Only accurately extracting the different target, the buildings, the trees, and the roads from the surface information, the three-dimensional can be constructed. In recent years, domestic and foreign has carried out extensive research for the trees extracted from the LIDAR data. But most of them aim at the forest area and the urban area is relatively few. In urban modeling, trees are an essential. Therefore, it is meaningful to research extraction of the trees in urban area.This paper uses the method of marked point process to directly extract the tree crown from the point cloud data and the method is an object-oriented method. The main research contents are as follows:1、Introduce the research background and significance of this paper, and reviewing the relevant research status for the extraction of trees in the LIDAR point cloud data.2、The theoretical knowledge involved in airborne LIDAR technology is discussed in detail, including the composition of the system, the main application areas as well as the correlation technology.3、Briefly describe the point process and the marked point process theory. According to the crown structure characteristics in the LIDAR point cloud data, establish Gibbs energy model, and put the target corresponding to the energy model so that can accurately extract the target.4、Use RJMCMC and simulated annealing algorithm to obtain the global optimal solution of the model.5、Introduce the sources of experimental data used in this paper. Generate point cloud feature image using the point cloud data, finally according to the algorithm proposed in this paper to extract the crown.