The　traditional　projections　onto　convex　sets　super-resolution　image　reconstruction　algorithm　leads　to　the　halo　effect　in　reconstructed　high　resolution　image,　so　we　present　a　super-resolution　reconstruction　method　based　on　wavelet　bicubic　interpolation　algorithm.　The　new　algorithm　utilizes　wavelet　bi-cubic　interpolation　algorithm　to　obtain　the　initial　image　of　POCS　algorithm　and　introduces　bilateral　filter　to　estimate　point　spread　function.　Bilateral　filtering　makes　non-edge　pixel　having　minimum　influence　on　the　edge　region.　Algorithm　also　uses　adaptive　relaxation　parameters　to　reduce　the　impact　of　error　motion　information.　Compared　with　traditional　algorithm,　experiment　results　show　that　the　proposed　reconstruction　algorithm　eliminates　the　halo　effect　obviously,　thus　the　reconstructed　image　can　achieve　a　good　visual　effect.　©　2011　IEEE.