用微信扫码二维码

分享至好友和朋友圈

WE ARE COMMITTED TO REPORTING THE LATEST FORESTRY ACADEMIC ACHIEVEMENTS

谯程骏, 聂丛乐, 韩春坛, 等. 基于实测高光谱数据的川西地区常见灌木树种识别[J/OL]. 四川林业科技, 2024, 45[2024-07-11]. DOI: 10.12172/202404060001
引用本文: 谯程骏, 聂丛乐, 韩春坛, 等. 基于实测高光谱数据的川西地区常见灌木树种识别[J/OL]. 四川林业科技, 2024, 45[2024-07-11]. DOI: 10.12172/202404060001
QIAO C J, NIE C L, HAN C T, et al. Identification of common shrub species in Western Sichuan based on measured hyperspectral data[J/OL]. Journal of Sichuan Forestry Science and Technology, 2024, 45[2024-07-11]. DOI: 10.12172/202404060001
Citation: QIAO C J, NIE C L, HAN C T, et al. Identification of common shrub species in Western Sichuan based on measured hyperspectral data[J/OL]. Journal of Sichuan Forestry Science and Technology, 2024, 45[2024-07-11]. DOI: 10.12172/202404060001

基于实测高光谱数据的川西地区常见灌木树种识别

Identification of common shrub species in Western Sichuan based on measured hyperspectral data

  • 摘要: 采用FieldSpec4 Hi-Res便携式地物光谱仪,对四川西部4种常见灌木树种高山柳(Salix cupularis)、金露梅(Potentilla fruticosa)、鲜卑花(Sibiraea angustata)、鲜黄小檗(Berberis diaphana Maxin)进行了野外光谱数据采集,对其原始光谱数据、一阶导数光谱数据、光谱去包络线等数据进行分析,提取“绿峰位置”、“绿峰幅值”、 “红谷位置”、 “红谷幅值”、 “红边位置”、 “红边幅值”、 “红边面积”、 “吸收峰面积”、 “吸收峰对称度”等光谱特征参量,用于神经网络树种识别模型的输入,对灌木树种进行分类和识别。神经网络模型的训练精度为:93.0%,验证精度为:83.3%,测试精度为:87.5%;模型精度较高,分类结果总体上较好。分类结果显示,高山柳和金露梅的精度最高;鲜卑花的训练精度较高,而测试精度较低,分类精度较不稳定,主要错分为金露梅;鲜黄小檗主要错分到金露梅。基于实测高光谱数据,建立神经网络树种识别模型,能有效区分不同灌木树种,为高光谱遥感开展森林资源树种分类和动态监测等提供理论和技术支持。

     

    Abstract: Using a FieldSpec4 Hi-Res portable land cover spectrometer, four common shrub species in western Sichuan, including Salix cupularis、Potentilla fruticosa、Sibiraea angustata and Berberis diaphana Maxin were collected for field spectral data. The original spectral, first-order derivative spectral, and spectral de envelope lines data were analyzed to extract "green peak position", "green peak amplitude", "red valley position", "red valley amplitude", "red edge position", "red edge amplitude", and "red edge amplitude", which are used as inputs for neural network tree species recognition models to classify and recognize shrub species. The training accuracy of the neural network model is 93.0%, the validation accuracy is 83.3%, and the testing accuracy is 87.5%. The model has high accuracy and overall good classification results. The classification results show that Salix cupularis and Potentilla fruticosa have the highest accuracy; The training accuracy of Berberis diaphana Maxin is high, while the testing accuracy is low, and the classification accuracy is unstable, mainly misclassified as Potentilla fruticosa; Sibiraea angustata is mainly misclassified as Potentilla fruticosa. Based on measured hyperspectral data, establishing a neural network tree species recognition model can effectively distinguish different shrub species, providing theoretical and technical support for hyperspectral remote sensing to carry out forest resource tree species classification and dynamic monitoring.

     

/

返回文章
返回