|
|
|
Cloud masking and fire detection algorithms using NOAA-AVHRR data- a case study over central Himalayan Region
Yogesh Kant, Gopal Biswas, V.
Krishna Prasad, K. V. S. Badarinath and S. K. Bhan
National Remote
Sensing Agency, Department of Space, Government of India, Balanagar - 500 037
Biomass burning associated with trace gas emissions is an
important source of trace gas emissions. The major cause of uncertainty in
determining the contribution of biomass combustion to trace gas emissions and
their environmental impact is due to lack of information on the extent of area
burnt, intensity of fires along with the biomass characteristics. Fire detection
through conventional means suffers from practical problems and it has been well
recognized that satellite data can be used effectively for detection of active
fires and scars from the biomass burning plumes. NOAA AVHRR satellite data has
five spectral channels (mm) viz., 0.58-0.68 (visible), 0.72-1.10(near infrared),
3.55-3.93 (mid-infrared), 10.3-11.3 (thermal infrared) and 11.5-12.5 (thermal
infrared). Although, there have been studies on forest fire detection using
thermal channels, the impact of clouds has been neglected. In the present study,
an attempt has been made to test one of the fire detection algorithms through
the combined use of middle and thermal infrared channels of NOAA-AVHRR satellite
data. Two cloud screening algorithms have been attempted in the study. In the
first method, a cloud-masking filter using channel five and channel one has been
employed. In the second method, from the channel four, using a threshold value,
the contaminating pixels are eliminated. Fire detection has been done using the
brightness temperature difference between the channel three and channels four
using the reference temperature difference threshold value. The methodology has
been applied on AVHRR data over Central Himalyan forest area where large
devastating fire occurred during May/June, 1995 affecting nearly 4,869 sq. km
area. Using the above data, cloud screening along with fire detection algorithms
have been implemented and a map showing the extent of active fire zones has been
generated. Results of the study suggest the potential use of AVHRR satellite
data in monitoring the fire events.
|
|
|