中文摘要：

将机载激光扫描点云数据转换成点云特征影像,提出基于多标记点过程的建筑物和树冠目标几何对象的自动提取方法。首先根据目标的几何特征建立吉布斯自由能变模型,通过目标的一致性建立该模型的数据项,通过目标的拓扑性质等空间特性建立该模型的先验项。然后利用可逆跳转马尔可夫蒙特卡洛（RJMCMC）算法进行采样,并采用模拟退火算法进行优化求解,实现建筑物目标和树冠目标几何对象的多目标自动提取。试验结果表明该方法能够从机载激光扫描数据中有效地提取建筑物和树冠,具有较强的稳健型。

英文摘要：

A multi-marked point process based method is used to extract building and tree crown from point cloud feature image generated from airborne LiDAR data.Firstly,the Gibbs free energy change model is build according to the geometric feature of the object.This model contains both a data coherence term which fits the objects to the data and a prior term which incorporates the prior knowledge of the object geometric properties.Then the previously defined model will be sampled using a RJMCMC（reverse jump Markov chain Monte Carlo） algorithm and optimized by a simulated annealing algorithm.The experimental results show that this method is capable of efficiently extracting building and tree crown from airborne LiDAR data with great robustness.