@article{fdi:010061471,
  title = {{E}stimating deforestation in tropical humid and dry forests in {M}adagascar from 2000 to 2010 using multi-date {L}and sat satellite images and the random farests classifier},
  author = {{G}rinand, {C}. and {R}akotomalala, {F}. and {G}ond, {V}. and {V}audry, {R}. and {B}ernoux, {M}artial and {V}ieilledent, {G}.},
  editor = {},
  language = {{ENG}},
  abstract = {{H}igh resolution and low uncertainty deforestation maps covering large spatial areas in tropical countries are needed to plan efficient forest conservation and management programs such as {REDD}+ ({R}educing {E}missions from {D}eforestation and {F}orest {D}egradation). {U}sing an open-source free software ({R}, {GRASS} and {QG}is) and an original statistical approach combining multi-date land cover observations based on {L}andsat satellite images and the random forests classifier, we obtained up-to-date deforestation maps for the periods 2000-2005 and 2005-2010 with a minimum mapping unit of 036 ha for 7.7 {M} hectares, i.e. 40.3% of the tropical humid forest and 20.6% of the tropical dry forest in {M}adagascar. {U}ncertainty in deforestation on the maps was calculated by comparing the results of the classification to more than 30,000 visual interpretation points on a regular grid. {W}e assessed accuracy on a per-pixel basis (confusion matrix) and by measuring the relative surface difference between wall-to-wall approach and point sampling. {A}t the pixel level, user accuracy was 84.7% for stable land cover and 60.7% for land cover change. {O}n average for the whole study area, we obtained a relative difference of 2% for stable land cover categories and 21.1% land cover change categories respectively between the wall-to-wall and the point sampling approach. {D}epending on the study area, our conservative assessment of annual deforestation rates ranged from 0.93 to 233%.yr(-1) for the humid forest and from 0.46 to 1.17%.yr(-1) for the dry forest. {H}ere we describe an approach to obtain deforestation maps with reliable uncertainty estimates that can be transposed to other regions in the tropical world.},
  keywords = {{D}eforestation ; {C}hange detection ; {C}lassification ; {L}and cover ; {L}andsat {TM} ; {M}achine learning ; {M}adagascar ; {R}andom forests ; {REDD} ; {MADAGASCAR}},
  booktitle = {},
  journal = {{R}emote {S}ensing of {E}nvironment},
  volume = {139},
  numero = {},
  pages = {68--80},
  ISSN = {0034-4257},
  year = {2013},
  DOI = {10.1016/j.rse.2013.07.008},
  URL = {https://www.documentation.ird.fr/hor/fdi:010061471},
}