Thanks for all your kind help. Here I make a summary for this question. The programs can do this are listed below.

1. PASS. (but it is fail to find the channel in my case).
2. Pymol with Caver scripts. (also fail to find the channel in my case).
3. Klaus Futterer and Hari jayaram suggest the program HOLE and it can accessed from http://d2o.biop.ox.ac.uk:38080/
4. Walter Novak suggests CASTp. It can be accessed form http://sts-fw.bioengr.uic.edu/castp/ or http://www.cgl.ucsf.edu/chimera/.
5. Beth A Wurzburg, Franz Gruswitz and Karsten Schaefer suggest the program MOLE. It can be accessed form http://troll.chemi.muni.cz/whitezone/development/mole/online/moleonline1.3/ or http://mole.chemi.muni.cz/download.php.
6. Franz Gruswitz also suggests another program HOLLOW which be accessed from http://hollow.sourceforge.net/.
7. Johansso suggests the program O can this with following protocol
   1/ make object of solvents in the tunnel, sel_off, sel_prop, sel_vis, sphere centred on tunnel...
   2/ mask from these solvents, ncs_mask_sph, ncs_mask_lay
   3/ asa of just protein
   4/ copy space group info from an electron density map to the asa 'map'
   5/ average the asa map, using the mask, and IDENT operator, no expansion gives a new map with a surface just around the tunnel.

Hope it will be helpful.
Thanks and best wishes.

--
Jiamu Du, Ph.D.
State Key Laboratory of Molecular Biology
Institute of Biochemistry and Cell Biology
Shanghai Institutes for Biological Sciences
Chinese Academy of Sciences
320 Yue-Yang Road
Shanghai 200031
P. R. China
Tel: +86-21-5492-1117
E-mail: [log in to unmask]