Global Vision Media Focus
The company was featured on CCTV a total of 15 times, and its reach through Party media, central media, and local official media reached a total of 50 million people.
Vegetation species identification based on narrowband anthocyanin index
Background Introduction
By utilizing the photosynthetic pigment index in a narrow band of hyperspectral imaging, vegetation species can be accurately distinguished. This case study uses hyperspectral data from the Xiguang-1 05 satellite (also known as Tianxianpei) and Sentinel-2 data to calculate the ARI1 anthocyanin index and extract forest land within farmland areas. Thanks to the 9 nm spectral resolution and narrow band design of the Xiguang-1 05 satellite (Tianxianpei), the reflectance difference at precise locations of 550 nm and 700 nm is significantly amplified, making forest land characteristics readily apparent. Compared to Sentinel-2 results, there are fewer false positives and the extracted results have less noise. Furthermore, a single image is sufficient, eliminating the need for time-series data, thus providing rapid and reliable spectral evidence for monitoring the "non-agriculturalization" of farmland.
Methods and Principles
Table 1. Comparison of Band Parameters
The leaf pigment index (ARI1) is used to measure stress-related pigments in vegetation. ARI1 is highly sensitive to anthocyanins in leaves; a higher ARI1 value indicates canopy growth or death. The calculation formula is as follows:
The value ranges from 0 to 0.2+, while the range for general green vegetation areas is 0.001 to 0.1.
As a narrow-band vegetation index, the closer the center wavelength is to the wavelength required by the formula and the smaller the FWHM, the better the index calculation results can reflect the true anthocyanin concentration.
We compared data from Sentinel-2 and Xiguang-1 05 satellite (the "Tianxianpei" satellite). The Sentinel-2 data was collected from April 1st to April 10th, 2025, and is a median composite result. The Xiguang-1 05 satellite (the "Tianxianpei" satellite) data was collected on April 6th, 2025.
Using data from Xiguang-1 05 satellite (Tianxianpei) and Sentinel-2, the two bands closest to 550nm and 700nm were selected, and the ARI1 index was calculated. For the index results, simple threshold segmentation was used to screen out woodlands in farmland, and the results of Xiguang-1 05 satellite (Tianxianpei) and Sentinel-2 were compared and evaluated.
Table 2. Comparison of wavebands used in index calculation
Results Display
Area 1
Figure 1. ARI1 index of Sentinel-2, Region 1
Figure 2. Farmland and woodland extraction results from Sentinel-2, Region 1
Figure 3. ARI1 index of Xiguang-1 05 satellite (Tianxianpei), region 1
Figure 4. Farmland and forest land extraction results from Xiguang-1 05 satellite (Tianxianpei), Region 1
Area 2
Figure 5. Farmland and woodland extraction results from Sentinel-2, Region 2
Figure 6. Farmland and woodland extraction results from Xiguang-1 05 satellite (Tianxianpei), Region 2
Area 3
Figure 7. Farmland and woodland extraction results from Sentinel-2, Region 3
Figure 8. Farmland and woodland extraction results from Xiguang-1 05 satellite (Tianxianpei), Region 3
The comparison results show that, compared with the ordinary NDVI index, the ARI1 index can reflect the differences between crops and can support application scenarios such as crop stress analysis and extraction of illegal forest land in farmland.
Because Sentinel-2 has a limited number of bands, the bands do not perfectly match the wavelengths required by the formula. Furthermore, Sentinel-2's wide bandwidth makes it impossible to accurately obtain the reflectance at a specific wavelength. Therefore, the range of values for Sentinel-2's ARI index calculation is shorter than that of the Xiguang-1 05 satellite (also known as the Tianxianpei satellite), indicating that the crop differentiation ability based on the ARI1 index is lower than that of the Xiguang-1 05 satellite (also known as the Tianxianpei satellite).
Based on the extraction results of farmland and woodland, using the same simple threshold segmentation method, Sentinel-2's extraction results have high salt-and-pepper noise, and there are more false positives in the western region of this image. In contrast, the results from Xiguang-1 05 satellite (also known as Tianxianpei) are better.
Compared to wideband multispectral data, hyperspectral data has a clear advantage in calculating narrowband vegetation indices.