Sliding　inspections　at　heights　and　in　harsh　environments　generally　pose　significant　challenges　for　the　autonomous　operation　of　unmanned　aerial　manipulator　(UAM).　This　article　investigates　the　issue　of　reliable　physical　interaction　by　UAM　for　sliding　inspections　under　uncertain　disturbances.　First,　a　second-order　image　dynamics　model　is　derived　using　the　virtual　camera　image　moment　to　provide　a　foundation　for　autonomous　positioning.　Next,　a　sliding-mode　disturbance　observer　is　designed　to　estimate　the　uncertain　disturbances　from　environments.　Then,　a　hybrid　image-based　compliant　control　strategy　is　proposed　to　meet　the　physical　interaction　requirements　of　UAM.　Particularly,　a　visual　positioning　controller　is　adopted　to　maintain　precise　positions　of　UAM,　while　a　visual　impedance　control　method　is　introduced　to　achieve　reliable　force　tracking　performance　in　the　contact　direction.　The　stability　of　the　proposed　strategy　is　analysed　through　Lyapunov　theory.　Finally,　physical　in-loop　implementation　and　real-world　experimental　studies　are　conducted　to　validate　the　feasibility　and　performance　of　the　developed　method,　demonstrating　accurate　tracking　performances　in　both　position　and　force　and　thereby,　exhibiting　the　potential　of　proposed　approach　to　work　for　sliding　inspection　effectively.