In　order　to　find　the　defects　in　ferromagnetic　materials,　a　non-contact　harmonic　detection　method　is　proposed.　According　to　the　principle　of　frequency　modulated　carrier　wave,　a　tunnel　magneto　resistance　harmonic　focusing　vector　array　detector　was　designed　which　radiates　lower　and　higher　frequency　electromagnetic　waves　through　the　coil　array　to　the　detection　targets.　We　use　bistable　stochastic　resonance　to　enhance　the　energy　of　collected　weak　target　signal　and　apply　quantum　computation　and　a　Sobol　low　deviation　sequence　to　improve　genetic　algorithm　performance.　Then　we　use　the　orthogonal　phase-locked　loop　to　eliminate　the　intrinsic　background　excitation　field　and　tensor　calculations　to　fuse　the　vector　array　signal.　The　finite　element　model　of　array　detector　and　the　magnetic　dipole　harmonic　numerical　model　were　also　established.　The　simulation　results　show　that　the　target　signal　can　be　identified　effectively,　its　focusing　performance　is　improved　by　2　times,　and　the　average　signal-to-noise　ratio　is　improved　by　9.6　times　after　the　algorithm　processing.　For　the　experiments,　we　take　Q235　steel　pipeline　as　the　object　to　realize　the　recognition　of　three　defects.　Compared　with　the　traditional　methods,　the　proposed　method　is　more　effective　for　ferromagnetic　materials　defects　detection.　©　2021.　The　Author(s).