Envir ONF — Total Terrier Surface

[link to description pdf](https://carto.ain.fr/webapps/externe/doc/ONF_notice_modeles_LiDAR_v2.pdf)

1 — Modeling principles

The present models are constructed from the detection of the apex identified on the digital height model from the LiDAR covers. Only heights greater than 13 meters associated with bowls of more than 5 square meters are preserved. Treatment shall be carried out excluding the ‘unwooded’ type of the IFN stand mapping v.2. Each apex is classified as hardwood or soft based on the slopes of its hoop and its reflectance measured in SPOT 6 images. The diameters at 1.30 m and the increments are estimated by the characteristics of the bowl associated with the apex. At the pixel scale of 26.5 m on the side, the earth surfaces predicted for each tree are summed by species and categories of diameters. The results obtained are corrected according to linear models. Models applied to apex and pixels are calibrated using the forest observatory plots. They are validated on a new set of 240 plots of 15 meters radius measured in 2015 and independent of the calibration set.

2 — Validity of modeled parameters

Generally, the validity of modeled parameters is not known for rare and atypical stands at the Ain Mountain scale. These stands are, for example, the plantations of larchs, the pine groves, the young leafy young leafy thinned... The validity in stands of black pine forests or sylvestres has not been verified. All data are for a pre-counting diameter of 17.5 cm to 1.30 m (unless otherwise stated).

3 — Precautions for the realisation of calculations based on the data of the rasters

For basal area, production, stem density, stand types, zero values shall be used in the absence of data. These cases correspond either to stands less than 13 metres high (a priori composed of trees with a diameter of less than 17.5 centimeters), or to a type of non-wooded IFN stand. For stand types, the zero value is a non-pre-accountable stand, or an unforested area. For height, diameter and Gini values, the NULL value shall be used in the absence of data. The edge pixels incorporate on a part of their surface an unwooded area. In case the extraction area incorporates such pixels, it is important to apply to the extraction surface a negative buffer that will exclude its points.

[数据集描述PDF链接](https://carto.ain.fr/webapps/externe/doc/ONF_notice_modeles_LiDAR_v2.pdf)

1 — 建模原理
本系列模型基于激光雷达（LiDAR）覆盖区域的数字高程模型上识别出的树冠顶点构建而成。仅保留高度大于13米且树冠投影面积超过5平方米的对象。处理过程需排除法国国家森林调查（IFN）第二代林分制图中的“无林地”类型。每个树冠顶点将依据其树冠轮廓的坡度与SPOT 6卫星影像获取的反射率，被划分为阔叶树或针叶树。树高1.30米处的胸径及其生长量，将通过与该顶点关联的树冠投影特征进行估算。在边长为26.5米的像素尺度下，单棵树木的预测地表面积将按树种与胸径等级进行汇总。所得结果将通过线性模型进行校正。针对树冠顶点与像素的模型，将借助森林观测样地进行校准。模型将在2015年实测的、独立于校准集的240个半径15米的新样地中进行验证。

2 — 建模参数的有效性
总体而言，在艾因山（Ain Mountain）尺度下，稀有或非典型林分的建模参数有效性尚未明确。此类林分例如落叶松人工林、松树林分、经间伐的幼龄阔叶树林……黑松林或欧洲赤松林的参数有效性尚未得到验证。除非另有说明，所有数据均对应胸高（1.30米处）预计数直径为17.5厘米的林木。

3 — 基于栅格数据开展计算的注意事项
对于断面积、生产量、茎干密度及林分类型，若缺少数据则需以零值填充。此类无数据情况要么对应高度不足13米的林分（理论上由胸径小于17.5厘米的林木组成），要么对应IFN林分制图中的非林地类型。就林分类型而言，零值代表未预计数林分或无林地区域。对于树高、胸径及基尼系数（Gini）值，若缺少数据则需以空值（NULL）填充。边缘像素的部分区域包含无林地。若提取区域包含此类像素，则需对提取区域应用负缓冲区以排除其中的无效点。