The MODIS contextual algorithm is composed of three basic parts, including preliminary thresholds to identify potential fire pixels, contextual tests to confirm fires among the potential fire pixels (Martin et al., 1999), and thresholds to reject false alarms. In the first part, the selection of fixed thresholds is subtle as an over-high setting runs a risk of omitting fire pixels. Meanwhile, an over-low setting causes more noise in deriving the parameters of the background pixels (Li et al., 2001) and generates more false alarms. The MODIS version 4 contextual algorithm employs fixed thresholds globally to identify potential fire pixels. For global applications, the preliminary thresholds cannot be set low enough to detect most small fires that can be physically detected for regional concern. Therefore, it needs improvement for fire monitoring and management at the regional scale.
The MODIS contextual algorithm is composed of three basic parts, including preliminary thresholds to identify potential fire pixels, contextual tests to confirm fires among the potential fire pixels (Martin et al., 1999), and thresholds to reject false alarms. In the first part, the selection of fixed thresholds is subtle as an over-high setting runs a risk of omitting fire pixels. Meanwhile, an over-low setting causes more noise in deriving the parameters of the background pixels (Li et al., 2001) and generates more false alarms. The MODIS version 4 contextual algorithm employs fixed thresholds globally to identify potential fire pixels. For global applications, the preliminary thresholds cannot be set low enough to detect most small fires that can be physically detected for regional concern. Therefore, it needs improvement for fire monitoring and management at the regional scale.
การแปล กรุณารอสักครู่..
