| Estimation of soil salinity and pH value based on surface hyperspectral and OLI images |
|
View Fulltext View/Add Comment Download reader |
| |
| DOI：10.7606/j.issn.1000-7601.2021.01.22 |
| Key Words: saline\|alkali soil salinity pH value vegetation index sensitive band hyper\|spectral Landsat 8 OLI image |
| Author Name | Affiliation | | SUN Yuan | College of Resources and Environmental Science, Ningxia University, Yinchuan, Ningxia 750021, China | | JIA Pingping | College of Resources and Environmental Science, Ningxia University, Yinchuan, Ningxia 750021, China | | SHANG Tianhao | College of Resources and Environmental Science, Ningxia University, Yinchuan, Ningxia 750021, China | | ZHANG Junhua | Institute of Environmental Engineering, Ningxia University, Yinchuan, Ningxia750021, China |
|
| Hits: 874 |
| Download times: 186 |
| Abstract: |
| The purpose of this paper is to improve the precision of salinity and alkalinity monitoring model with measured vegetation canopy spectrum and Landsat 8 OLI multi\|spectral image on Northern Yinchuan Plain of Ningxia. In this paper, the authors took six transformations on the measured vegetation canopy hyperspectral and image multi\|spectral reflectance, chose the best spectral transformation form, the most sensitive vegetation index and bands to establish the soil salt content and pH value estimation model using the resampled actual measurement data and corrected Landsat 8 OLI image inversion of soil salinity and alkalinity. The results showed that the measured vegetation hyperspectral EVI model established by inverse logarithmic transformation and the measured vegetation hyperspectral model established by smoothed sensitive bands have higher accuracy in estimating soil pH value, and the model determination coefficients are 0.6257 and 0.5975, respectively. The measured hyperspectral vegetation index and sensitive bands corrected the salt content and pH estimation models of the Landsat 8 OLI image, respectively. The coefficients of determination of the sensitive vegetation index and the salt content model of the sensitive bands were increased by 0.3207 and 0.3762, respectively. The model determination coefficients were increased by 0.2065 and 0.2487, respectively. The study used sensitive vegetation index and sensitive bands to estimate soil salinity and pH value simultaneously, and realized the scale transformation of the soil salt content and pH value spectral estimation model from field measurements of spectral scales to spectral scale of multi\|spectral remote sensing, and the results could provide a theoretical reference for further improvement of the accuracy of quantitative remote sensing monitoring of soil salt content and pH value at the local and similar regions. |
|
|
|