2.3.1. Original Band Information

Then, the band information was extracted from the processed images using the vector file of the ground sample. Eight bands (Table 2) of the Sentinel-2 data were selected for this study [30–32], excluding bands 1, 9, 10, 11, and 12 because these bands are mainly associated with the atmosphere or water vapor.


**Table 2.** Selected band information of Sentinel-2.

#### 2.3.2. Traditional Vegetation Indices

The potential of six traditional vegetation indices for estimating FSV, calculated from the band reflectance extracted from the Sentinel-2 data (Table 3) were initially tested. Normalized Difference Vegetation Index (NDVI) reflects the background influence of plant canopy and is concerned with vegetation coverage. It is a vegetation index frequently utilized for detecting the growth status of plants. The difference vegetation index (DVI) can also reflect changes in vegetation coverage very well, and within a certain range of vegetation coverage, the DVI rises with the growth of biomass. The ratio vegetation Index (RVI) is a highly sensitive indicator parameter for monitoring green plants, which can be used to detect vegetation status and estimate the FSV. This index is the ratio of light scattered in the near-infrared to light absorbed in the red band, which lessens the effect of the atmosphere and terrain. The perpendicular vegetation index (PVI) represents the vertical distance from the vegetation pixel to the soil brightness line in the two-dimensional coordinate system of R—NIR and is less sensitive to the atmosphere than other vegetation indices. The transformed vegetation index (TVI) is based on the NDVI and introduces a constant of 0.5 to convert the negative value that the NDVI may take into a positive value. The EVI not only inherits the advantages of the NDVI, but also improves the saturation of high vegetation areas, incomplete correction of atmospheric effects, and soil background. The enhanced vegetation index (EVI) can improve the sensitivity of vegetation in high biomass areas and reduce the influence of soil background and atmosphere.

**Table 3.** Several traditional vegetation indices calculated based on Sentinel-2 data.


2.3.3. Novel Vegetation Index Based on Red-Edge Bands

The accuracy of traditional vegetation indices to estimate FSV is severely affected by the light saturation phenomenon. While the three red-edge bands in the Sentinel-2 data have been proven to be an effective way to improve the estimation of the forest parameters, unfortunately only one or two of the red-edge bands were used in existing indices. Therefore, to maximize the ability to estimate FSV using the three red-edge bands in the Sentinel-2 data, a novel vegetation index based on existing NDVI construction principles, the 4-band red-edge NDVI (*NDVIRE*), such as Formula (1) was developed. According to the novel index construction rules, as elaborated in previous studies, in the *NDVIRE* formula, instead of using the NIR band, the reflectance values of *RE3* and *RE2* are averaged using weights and are substituted. Similarly, the Red band is replaced with a weighted average of the reflectance values of *RE1* and *RE2* [27]. The weighting coefficients "*α*" and "*β*" are designed to define the optimal proportion of each band in the construction of the novel index.

$$NDVI\_{RE} = \frac{\left(\alpha \cdot R\_{RE3} + (1 - \alpha) \cdot R\_{RE2}\right) - \left(\beta \cdot R\_{Red} + (1 - \beta) \cdot R\_{RE1}\right)}{\left(\alpha \cdot R\_{RE3} + (1 - \alpha) \cdot R\_{RE2}\right) + \left(\beta \cdot R\_{Red} + (1 - \beta) \cdot R\_{RE1}\right)}\tag{1}$$

where *RRE*1, *RRE*2, *RRE*3, and *RRed* are the reflectance of B5, B6, B7, and B4, respectively. "*α*" and "*β*" represent weighting coefficients. The value range of "*α*" and "*β*" is (0,1), and the step size is 0.1.
