Aiming　at　the　problems　of　limited　single　image　feature　information　and　insufficient　algorithm　recognition　ability　in　traditional　power　quality　disturbance　(PQD)　recognition　schemes,　a　PQD　recognition　method　based　on　feature　image　combination　and　modified　ResNet-18　is　proposed　according　to　the　idea　of　feature　fusion.　First,　a　series　of　intrinsic　mode　functions　(IMFs)　and　residual　components　are　obtained　by　variational　mode　decomposition　(VMD)　of　PQD　signals.　Then,　the　IMFs,　residual　components,　original　disturbance　signals　and　Subtract　components　are　longitudinally　spliced　into　component　matrix,　and　the　signal-image　conversion　method　is　used　to　generate　the　feature　component　color　map.　Meanwhile,　continuous　wavelet　transform　(CWT)　is　performed　on　the　original　disturbance　signal　to　generate　the　wavelet　time-frequency　diagram.　Finally,　the　feature　component　color　map　and　wavelet　time-frequency　diagram　are　combinatorically　input　into　the　modified　six-channel　ResNet-18　training　and　the　learning　on　how　to　recognize　the　PQD.　The　PQD　recognition　method　is　analyzed　through　simulation　and　compared　with　the　commonly　used　recognition　system.　The　results　show　that　the　proposed　method　has　good　anti-noise　performance　and　can　better　extract　the　PQD　feature　information　to　achieve　higher　recognition　accuracy.　©2024　Chin.Soc.for　Elec.Eng.　2531.