Channel rendezvous is a prerequisite in cognitive radio networks, where secondary users choose a common available channel and establish a connection. It is difficult to guarantee rendezvous in heterogeneous cognitive radio networks, where no universal channel set is shared among secondary users. In this paper, we propose two Galois-field-based channel hopping (GFCH) heterogeneous rendezvous algorithms: one is global-channel-oriented and the other is local-channel-oriented. The algorithms generate periodic channel-hopping sequences by performing the operations of plus and multiply in the Galois field. The sequence in each period includes three phases: access all the channels in a clockwise direction, access the parity channel, and access all the channels in a counterclockwise direction. As in a distributed environment, global channels may not be shared among secondary users; we then propose a local Galois-field-based channel hopping (L-GFCH) rendezvous algorithm utilizing only local available channels. Furthermore, we analyze the theoretical values of the maximum-time-to-rendezvous of GFCH and L-GFCH. Extensive simulations confirm that GFCH and L-GFCH outperform existing algorithms in terms of expected-time-to-rendezvous in global and local scenarios, respectively.