研究目的
To adjust models to estimate dendrometric characteristics of the Brazilian dry tropical forest (Caatinga) from Landsat 5 TM sensor data.
研究成果
The study successfully developed models to estimate tree height and wood volume using Landsat 5 TM data, with R2 values of 0.4 and 0.6 respectively. The red band, NDVI, and Savi were key explanatory variables. Remote sensing is a valuable tool for monitoring Caatinga forests, offering advantages in coverage and cost over conventional methods, despite some limitations in accuracy for certain variables.
研究不足
The basal area did not correlate significantly with explanatory variables, possibly due to the horizontal structure of Caatinga vegetation. The RMSE for volume estimation was high (42%), exceeding acceptable levels for forest inventories. The study used data from only two fragments and a single Landsat scene, limiting generalizability. Future work should consider other sensors and techniques like Lidar, and larger plot sizes.
1：Experimental Design and Method Selection:
The study used remote sensing data from Landsat 5 TM sensor to develop regression models for estimating dendrometric variables. Methods included radiometric and geometric corrections, extraction of spectral bands and vegetation indices, and multiple linear regression analysis.
2：Sample Selection and Data Sources:
Data were collected from 60 inventory plots (400 m2 each) in two municipalities in Sergipe, Brazil, measuring tree diameter and height. Remote sensing data came from a Landsat 5 TM scene.
3：List of Experimental Equipment and Materials:
Equipment included a telescopic pole for height measurement, a measuring tape for diameter, and a Garmin 12XL GPS for coordinates. Materials involved Landsat 5 TM imagery and software for data processing.
4：Experimental Procedures and Operational Workflow:
Field campaigns in June and July 2008 measured dendrometric variables. Image preprocessing included georeferencing and atmospheric correction using the 6S model. Spectral bands and indices (SR, NDVI, Savi) were extracted and correlated with field data.
5：Data Analysis Methods:
Statistical analysis involved correlation coefficients, multiple linear regression, t-tests, residual analysis, variance inflation factor (VIF) test, and cross-validation with R2 and RMSE metrics.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
