基于LiDAR和DOM数据的薇甘菊自动识别与分布预测

薇甘菊攀援能力强，生长速度快，对生态环境和生物多样性造成了严重威胁。卫星遥感数据是薇甘菊识别和预测的主要数据源，但现有的数据存在分辨率低、过境时间长和云层遮挡等方面的局限性，对薇甘菊识别和预测的精度较低，为此，提出了一种结合机载激光数据(LiDAR)和航摄多光谱数据(DOM)的薇甘菊爆发区域自动识别及入侵概率预测方法。采用面向对象的多尺度分割方法对研究区内薇甘菊爆发点进行自动识别，并利用林场内冠层高度模型、植被覆盖度、坡度、坡向等数据，采用Logistic回归方法对薇甘菊入侵分布概率进行预测。结果表明：面向对象的多尺度分割方法能较好地提取研究区内薇甘菊爆发区域，识别精度较好，错分率为4.66%，漏检率为0.41%；Logistic回归模型对薇甘菊的入侵分布概率有较好的预测效果，准确率为88.46%。该方法可实现大范围内薇甘菊的精确识别及预测，可服务于薇甘菊的综合防控与监测，为薇甘菊的入侵监测提供有力支撑。

Automatic Identification and Predictive Analysis of Mikania micrantha Based on LiDAR and DOM Data

Mikania micrantha has strong climbing ability and amazing growth speed, which poses a serious threat to the surrounding ecological environment and biodiversity. Satellite remote sensing data is the main data source for identification and prediction of Mikania micrantha. However, the existing data have limitations such as low resolution, long transit time and cloud shielding, and the accuracy of identification and prediction of Mikania micrantha is low. In view of this, a method for automatic identification of Mikania micrantha outbreak area and invasion probability prediction based on airborne laser data and aerial multispectral data was proposed.Object-oriented multi-scale segmentation method was used to automatically identify the outbreak points of Mikania micrantha in the study area, and Logistic regression method was used to predict the invasion distribution probability of Mikania micrantha by using canopy height model, vegetation coverage, slope and slope aspect data in the forest farm. The results showed that the object-oriented multi-scale segmentation method could extract the Mikania micrantha outbreak area in the study area, and the identification accuracy was high, the misclassification rate was 4.66%, and the missed detection rate was 0.41%.Logistic regression model had a good prediction effect on the invasion distribution probability of Mikania micrantha, and the correct rate was 88.46%.This method can realize accurate identification and prediction of Mikania micrantha in a wide range, and can serve for comprehensive prevention and control and monitoring of Mikania micrantha, providing strong support for invasion monitoring of Mikania micrantha.