Many　indoor　applications　require　spatial　information　about　objects.　Example　applications　include　item　finding,　object　level　mapping　and　large-scale　objects　managing　in　warehouses　or　libraries.　RFID　technology　enables　the　identification　and　location　of　objects　in　our　daily　life　by　integration　of　RFID　tags.　The　RF　signals　that　the　tags　backscatter　include　a　lot　of　information　about　the　environment　of　the　objects　which　can　be　used　to　derive　the　objects＇　spatial　information.　In　2016,　the　demand　for　applications　in　the　apparel　industry　alone　exceeded　4.6　billion　RFID　labels.　Due　to　the　emergence　of　three-dimensional　(3D)　printing,　Microsoft　and　Disney　Research　have　attempted　to　incorporate　RFID-like　embedded　encoding　information　directly　into　objects　during　the　manufacturing　process,　which　may　allow　objects　to　be　uniquely　identified　and　sensed　in　the　future.　To　this　end,　how　to　additionally　aware　the　precision　spatial　position　of　the　tagged　target　is　an　important　and　pressing　research　topic.　Traditional　localization　methods　based　on　RF　signals　with　backscatter　tags　can　be　classified　into　the　following　two　types:　(1)　trilateration　approaches　that　usually　require　multiple　RFID　readers　placed　in　the　environment,　where　the　target　position　is　determined　by　solving　nonlinear　equations　based　on　the　target-reader　geometric　relationship;　and　(2)　scene　analysis　approaches　that　rely　on　the　implementation　of　numerous　tags　with　known　positions　as　anchor　nodes.　However,　both　of　these　approaches　have　high　implementation　costs　and　it　is　difficult　to　apply　them　extensively　in　practice.　In　this　paper,　we　propose　a　3D　localization　scheme　based　on　phase　interferometric,　named　3DinSAR.　Our　localization　method　does　not　need　any　reference　tags　and　only　one　movable　reader　integrated　with　one　antenna　is　required　in　order　to　construct　the　synthetic　arrays　to　implement　the　locating　system.　The　main　idea　is　based　on　the　2D　naive　SAR　localization.　Rather　than　traversing　all　the　spatial　pixels　which　is　time　consuming　and　hence　not　efficient,　we　extend　the　naive　hologram　localization　into　a　3D　scheme　by　using　the　spatial　domain　phase　difference　which　relates　to　the　wave　path　difference　and　the　tags＇　height　as　additional　information.　Density-based　spatial　clustering　method　DBSCAN　is　used　to　give　the　final　estimation　target　location　to　improve　the　accuracy.　We　also　proposed　an　aperture-beam　predicting　method　ABP　to　improve　the　real-time　performance　of　2D　SAR　method.　In　summary,　the　mainly　contributions　of　this　paper　are:　(1)We　propose　a　3D　localization　scheme　based　on　phase　interferometric　for　passive　RFID　system,　named　3DinSAR,　which　is　able　to　precisely　estimate　the　tag＇s　3D　location.　(2)We　further　reduce　the　computational　time　required　by　using　an　aperture　beam　prediction　(ABP)　method,　in　order　to　improve　the　real-time　performance.　The　main　feature　of　the　ABP　method　involves　estimating　the　direction　of　arrival　in　order　to　ignore　any　unnecessary　pixels　when　creating　the　holographic　images.　(3)Experiments　are　performed　where　3DinSAR　is　implement　with　commercial　off-the-shelf　(COTS)　RFID　products.　Comparing　with　naive　2D　SAR　method,　the　simulation　results　demonstrate　ABP　method　has　discount　99.4%　of　the　time　consumption.　The　experimental　results　from　an　indoor　office　environment　show　a　spatial　median　error　of　24　cm　with　minimum　error　6cm　for　3DinSAR,　and　the　detection　time　maintain　within　95　ms.　©　2019,　Science　Press.　All　right　reserved.