One　important　way　to　extend　the　lifetime　of　wireless　sensor　networks　is　to　deploy　the　sensors　in　a　dense　manner.　The　redundancy　among　the　sensed　data　demonstrates　that　it　is　not　efficient　to　collect　raw　data　from　all　nodes　in　the　network　if　we　further　consider　that　the　data　is　generally　spatial-correlated.　The　node　scheduling　strategy　aims　at　selecting　a　set　of　representative　nodes　to　provide　the　required　data　service　in　a　periodic　manner　with　accuracy　guarantee.　This　strategy　can　effectively　reduce　the　number　of　active　nodes　and　the　amount　of　messages　in　the　network,　and　extend　the　network　lifetime　accordingly.　In　this　paper,　we　firstly　introduce　how　to　model　the　spatial　correlation　among　sensed　data　by　Markov　Random　Field　(MRF)　model.　Secondly,　we　formulate　the　problem　definitions,　namely,　the　data　amendment　problem　which　maximizes　the　data　coverage　range　for　a　given　node　by　amending　the　raw　noise-corrupted　data　from　the　neighbors,　while　the　representative　nodes　selection　problem　focuses　on　reducing　the　number　of　representative　nodes　and　covering　all　nodes　in　the　network,　and　the　node　scheduling　problem　aims　at　maximizing　the　network　lifetime.　Thirdly,　we　propose　a　novel　Data　Amendment　Procedure　(DAP),　Representative　node　Selection　Procedure　(RSP)　and　energy-efficient　Node　Scheduling　Algorithm　(NSA)　respectively　for　these　above　problems.　Finally,　extensive　experiments　demonstrate　that　the　proposed　node　scheduling　algorithm　can　significantly　improve　the　network　lifetime　compared　with　related　works　with　an　average　increment　of　about　80%.　(C)　2015　Elsevier　Inc.　All　rights　reserved.