Today we widely face the problem of spectrum scarcity and inefficiency in spectrum usage. A cognitive radio must detect the presence of any primary or secondary users in the frequen-cy bands to prevent interference. The primary task of a cognitive radio is to do a spectral estimation to locate the segments, which have white zones, and segments, which has grey zones. Once it is able to detect the white space, cognitive radio has to use this space to send signals at randomly located center frequencies with variable bandwidths. In this paper, we present a white space detection technique and an analysis channelizer based on a variant of the standard channelizer engine.