Spread spectrum signals are widely used for secure communications, as well as for multiple access interference (MAI). They have many interesting properties, including low probability of interception. Indeed, direct sequence spread spectrum systems (DSSSS) use a periodical pseudo-random sequence to modulate the baseband signal before transmission. Interception of DSSS signals is a difficult task due to the hiding of the PN sequence, especially if the accompanied noise is non-Gaussian and the channel is frequency selective. In this paper, a receiver is proposed to intercept asynchronous direct sequence code division multiple access (DS-CDMA) signals received from different users and estimate the PN sequences of each one of them. The signals are assumed to be transmitted over frequency selective fading channel and contaminated with Class-A impulsive noise. The proposed interception receiver is based on Independent Component Analysis (ICA), the pseudo random sequences is estimated from the eigenvector subspace in case of single path. In multipath situations, the estimation is the overlapped form of the pseudo random sequences, a triple correlation function (TCF) is introduced to eliminate the influence of multipath. The performance of proposed receiver is evaluated in terms of the number of correct estimated codes for m-sequence over typical urban channel in presence of impulsive noise. The results show that the interception receiver performs is 100% code estimation at -8dB signal to noise ratio.