One of the main difficult problems is how to choose interesting point efficiently in 3 dimensional stereo restoration. Although different algorithms are put forward in

these methods cannot restrain disturbance noise availably and are very rigorous on the illumination with large calculation. Based on these

a kind of method to fast extract feature point in 2 dimensional image is presented in this paper. It is proved that the algorithm mentioned in this paper has following remarkable advantages such as strong anti jamming capacity

powerful adaptability to outside illumination

simplicity and fast speed of computation

etc. after comparing the result of this algorithm with that of the algorithm in on extracting feature point by experiment. Moreover

it is testified that this kind of algorithm is valid after 3 dimensional restoration of standard piece with given 3 dimensional coordination. The accuracy of feature point extracted by the algorithm is high.