A　narrowband　tone　burst　excitation　is　usually　required　in　most　Lamb　wave　inspection　techniques　in　order　to　reduce　the　dispersion　effect　and　maintain　mode　purity,　especially　in　the　case　that　dispersion　compensation　is　hard　to　implement,　such　as　the　scenarios　prior　knowledge　of　accurate　dispersion　curves　is　not　available.　The　duration　of　the　excitation　is　relatively　large　because　of　its　narrow　bandwidth,　which　degrades　the　resolution　for　inspection.　Following　the　emerging　sparse　representation　theory,　a　combined　method　of　sparse　reconstruction　and　delay-and-sum　(DAS)　for　high　resolution　Lamb　wave　inspection　is　proposed　in　this　paper.　The　scattering　signals　are　compressed　in　time　domain　without　losing　the　information　of　the　arrival　time　of　each　echo.　Unlike　most　methods　that　signal　decomposition　and　reconstruction　use　the　same　dictionary,　the　proposed　approach　utilizes　two　different　dictionaries:　one　for　signal　decomposition　and　another　for　feature　extraction.　The　DAS　imaging　method　combined　with　the　feature　signals　is　applied　to　damage　imaging　for　carbon　fiber　reinforced　plastic　(CFRP)　laminates.　The　effectiveness　of　the　proposed　approach　is　validated　on　a　quasi-isotropic　laminated　CFRP　plate.　The　imaging　results　show　that　the　proposed　approach　can　achieve　a　higher　localization　resolution　and　lower　noise　floor　without　degrading　the　localization　accuracy.