Near-field-based　imaging　methods　have　been　widely　used　to　construct　images　of　complex　objects.　A　large　number　of　near-field　samples　are　needed　to　obtain　an　accurate　image,　resulting　in　high　implementation　expenditures.　By　utilizing　the　fact　that　the　electric　field　varies　fast　near　strong　sources,　this　paper　proposes　an　adaptive　sampling　strategy　to　add　near　field　sampling　points　near　the　strong　sources　adaptively.　With　the　proposed　strategy,　the　initial　uniform　sampling　is　conducted　to　obtain　rough　source　data.　Then　the　region　growing　method　is　applied　to　identify　multiple　strong　sources.　More　sampling　points　are　added　around　every　strong　source　point　to　extract　all　the　source　information　accurately.　The　experiment　results　demonstrate　that　the　proposed　strategy　can　effectively　reconstruct　radiation　sources　using　significantly　less　time　than　the　existing　methods.