With　the　development　of　information　technology,　the　computing　power　of　mobile　devices　continues　to　increase,　and　users　can　use　some　computationally　complex　programs　on　mobile　devices.　At　present,　research　related　to　pavement　recognition　is　mainly　in　the　field　of　unmanned　vehicles.　The　algorithm　proposed　in　this　paper　is　aimed　at　the　pavement　recognition　which　based　on　mobile　devices.　This　paper　presents　a　method　for　identifying　pedestrian　paths　in　green　spaces.　The　method　uses　image　filtering　and　morphological　processing　to　optimize　the　calculation　amount　of　the　image　which　reduce　the　operation　time　of　the　single　image,　and　classify　the　main　color　content　of　the　image.　And　then　process　the　image　in　the　HSV　color　space.　Finally,　we　propose　a　two-dimensional　matrix　template　for　extracting　the　pavement　edge　information　of　the　processed　image.　Experimental　results　show　that　this　method　can　quickly　process　images,　and　effectively　identify　the　pavement　edge　and　pavement　area　of　the　target　scene,　enabling　the　algorithm　to　be　applied　to　real-time　pavement　recognition　of　mobile　devices.　The　results　obtained　by　this　method　can　be　used　in　mobile　devices　to　provide　functions　such　as　road　recognition　or　augmented　reality　navigation.　©　2019　IEEE.