@article{fdi:010082634,
  title = {{I}nput imagery, classifiers, and cloud computing : insights from multi-temporal {LULC} mapping in the {C}ambodian {M}ekong {D}elta},
  author = {{O}rieschnig, {C}. {A}. and {B}elaud, {G}. and {V}enot, {J}ean-{P}hilippe and {M}assuel, {S}ylvain and {O}gilvie, {A}ndrew},
  editor = {},
  language = {{ENG}},
  abstract = {{T}he increased open-access availability of radar and optical satellite imagery has engendered numerous land use and land cover ({LULC}) analyses combining these data sources. {I}n parallel, cloud computing platforms have enabled a wider community to perform {LULC} classifications over long periods and large areas. {H}owever, an assessment of how the performance of classifiers available on these cloud platforms can be optimized for the use of multi-imagery data has been lacking for multi-temporal {LULC} approaches. {T}his study provides such an assessment for the supervised classifiers available on the open-access {G}oogle {E}arth {E}ngine platform: {N}aive {B}ayes ({NB}), {C}lassification and {R}egression {T}rees ({CART}), {R}andom {F}orest ({RF}), {G}radient {T}ree {B}oosting ({GTB}), and {S}upport {V}ector {M}achines ({SVM}). {A} multi-temporal {LULC} analysis using {S}entinel-1 and 2 is implemented for a study area in the {M}ekong {D}elta. {C}lassifier performance is compared for different combinations of input imagery, band sets, and training datasets. {T}he results show that {GTB} and {RF} yield the highest overall accuracies, at 94% and 93%. {C}ombining optical and radar imagery boosts classification accuracy for {CART}, {RF}, {GTB}, and {SVM} by 10-15 percentage points. {F}urthermore, it reduces the impact of limited training dataset quality for {RF}, {GTB}, and {SVM}.},
  keywords = {{CART} ; google earth engine ; gradient tree boosting ; {LULC} ; {R}andom {F}orest ; sentinel-1 and-2 ; {SVM} ; {CAMBODGE} ; {MEKONG} {DELTA}},
  booktitle = {},
  journal = {{E}uropean {J}ournal of {R}emote {S}ensing},
  volume = {54},
  numero = {1},
  pages = {398--416},
  year = {2021},
  DOI = {10.1080/22797254.2021.1948356},
  URL = {https://www.documentation.ird.fr/hor/fdi:010082634},
}