“…VI has been indicators of the vegetation [29][30][31]. The experiments would discuss whether they are contributive to the improvement of tree species classification.…”
Explicit information of tree species composition provides valuable materials for the management of forests and urban greenness. In recent years, scholars have employed multiple features in tree species classification, so as to identify them from different perspectives. Most studies use different features to classify the target tree species in a specific growth environment and evaluate the classification results. However, the data matching problems have not been discussed; besides, the contributions of different features and the performance of different classifiers have not been systematically compared. Remote sensing technology of the integrated sensors helps to realize the purpose with high time efficiency and low cost. Benefiting from an integrated system which simultaneously acquired the hyperspectral images, LiDAR waveform, and point clouds, this study made a systematic research on different features and classifiers in pixel-wised tree species classification. We extracted the crown height model (CHM) from the airborne LiDAR device and multiple features from the hyperspectral images, including Gabor textural features, gray-level co-occurrence matrix (GLCM) textural features, and vegetation indices. Different experimental schemes were tested at two study areas with different numbers and configurations of tree species. The experimental results demonstrated the effectiveness of Gabor textural features in specific tree species classification in both homogeneous and heterogeneous growing environments. The GLCM textural features did not improve the classification accuracy of tree species when being combined with spectral features. The CHM feature made more contributions to discriminating tree species than vegetation indices. Different classifiers exhibited similar performances, and support vector machine (SVM) produced the highest overall accuracy among all the classifiers.
“…VI has been indicators of the vegetation [29][30][31]. The experiments would discuss whether they are contributive to the improvement of tree species classification.…”
Explicit information of tree species composition provides valuable materials for the management of forests and urban greenness. In recent years, scholars have employed multiple features in tree species classification, so as to identify them from different perspectives. Most studies use different features to classify the target tree species in a specific growth environment and evaluate the classification results. However, the data matching problems have not been discussed; besides, the contributions of different features and the performance of different classifiers have not been systematically compared. Remote sensing technology of the integrated sensors helps to realize the purpose with high time efficiency and low cost. Benefiting from an integrated system which simultaneously acquired the hyperspectral images, LiDAR waveform, and point clouds, this study made a systematic research on different features and classifiers in pixel-wised tree species classification. We extracted the crown height model (CHM) from the airborne LiDAR device and multiple features from the hyperspectral images, including Gabor textural features, gray-level co-occurrence matrix (GLCM) textural features, and vegetation indices. Different experimental schemes were tested at two study areas with different numbers and configurations of tree species. The experimental results demonstrated the effectiveness of Gabor textural features in specific tree species classification in both homogeneous and heterogeneous growing environments. The GLCM textural features did not improve the classification accuracy of tree species when being combined with spectral features. The CHM feature made more contributions to discriminating tree species than vegetation indices. Different classifiers exhibited similar performances, and support vector machine (SVM) produced the highest overall accuracy among all the classifiers.
“…Harris et al (2011) concluded that the spectral index based on NBR and enhanced by LST (initially proposed by Veraverbeke et al, 2011) slightly outperformed NBR to assess burn severity in chaparral. Similarly, Zheng et al (2016) proposed a new index based on LST and enhanced vegetation index (EVI) and showed that it performed equally well for post-fire areas covered with both sparse vegetation and dense vegetation and relatively better than some commonly used burn severity indices. Quintano et al (2013) concluded that MESMA fraction images enable accurate burn severity mapping in Spanish Mediterranean ecosystems.…”
Section: U N C O R R E C T E D P R O O Fmentioning
Forest fires are incidents of great importance in Mediterranean environments. Landsat data have proven to be suitable for evaluating post-fire vegetation damage and determining different levels of burn severity, which is crucial for planning post-fire rehabilitation. This study assessed the utility of combined Multiple Endmember Spectral Mixture Analysis (MESMA) fraction images and Land Surface Temperature (LST) to accurately map burn severity. We studied a large convection-dominated wildfire, which occurred on 19-21 September 2012 in Spain, in a zone dominated by Pinus pinaster Ait. Burn severity degree (low, moderate, and high) was measured 2-3 months after fire in 111 field plots using the Composite Burn Index (CBI). Four fraction images were generated using MESMA from the reflective bands of a post-fire Landsat 7 Enhanced Thematic Mapper (ETM +) image: 1.-char, 2.-green vegetation (GV), 3.-non-photosynthetic vegetation and soil (NPVS) and 4.-shade. The thermal band was converted to LST using a single channel algorithm. Next, Multinomial Logistic Regression (MLR) was used to obtain the probability of each burn severity level from MESMA fraction images and LST. Finally, a burn severity map was generated from the probability images and independently validated using an error matrix, producer and user accuracies per class, and κ statistic. MLR identified the char fraction image and LST as the only significant explanatory variables when burn severity acted as the response variable. Two burn severity degrees (low-moderate and high) were finally considered to build the final burn severity map. In this way, we reached a higher accuracy (κ = 0.79) than using the original three burn severity levels (κ = 0.66). Our study demonstrates the validity of combining fraction images and LST from Landsat data to map burn severity accurately in Mediterranean countries.
“…Changes in land surface albedo (LSA) and land surface temperature (LST) have been used to estimate burn severity [45,[147][148][149]. Quintano et al [147] found that changes in LST showed high agreement with field measured CBI when used to map burn severity for an ecosystem dominated by maritime pine (Pinus pinaster) in Sierra del Teleno, Spain.…”
Wildfire plays an important role in ecosystem dynamics, land management, and global processes. Understanding the dynamics associated with wildfire, such as risks, spatial distribution, and effects is important for developing a clear understanding of its ecological influences. Remote sensing technologies provide a means to study fire ecology at multiple scales using an efficient and quantitative method. This paper provides a broad review of the applications of remote sensing techniques in fire ecology. Remote sensing applications related to fire risk mapping, fuel mapping, active fire detection, burned area estimates, burn severity assessment, and post-fire vegetation recovery monitoring are discussed. Emphasis is given to the roles of multispectral sensors, lidar, and emerging UAS technologies in mapping, analyzing, and monitoring various environmental properties related to fire activity. Examples of current and past research are provided, and future research trends are discussed. In general, remote sensing technologies provide a low-cost, multi-temporal means for conducting local, regional, and global-scale fire ecology research, and current research is rapidly evolving with the introduction of new technologies and techniques which are increasing accuracy and efficiency. Future research is anticipated to continue to build upon emerging technologies, improve current methods, and integrate novel approaches to analysis and classification.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.