Parking　cell　detection　is　one　of　the　key　technologies　in　parking　lot　monitoring　and　management　system,　this　paper　proposes　a　parking　cell　detection　algorithm　of　multiple　video　features,　in　view　of　complex　background　and　different　illumination　conditions　in　the　parking　lot.　Firstly,　this　paper　has　fully　used　the　geometrical　and　statistical　features　of　the　parking　cell　information,　and　extracted　the　features　parameter　of　parking　cell,　and　then　designed　three　kinds　of　different　solution　from　mathematics　modeling　angle.　Finally,　many　images,　which　are　regarded　as　the　parking　cell　detection　images　in　different　weather　and　conditions　have　been　tested　by　those　proposed　algorithms　continuously　for　14　days,　and　comparisons　have　been　made　with　the　results　in　the　proposed　algorithms.　The　experimental　result　indicated　that　the　algorithm　of　Bayes　methods　with　PCA　is　superior　to　other　algorithms　in　aspects　of　operation　time,　recognition　accuracy,　and　the　robustness.　Its　detection　rate　reaches　as　high　as　98.99%.