The deployment of highly dense wireless local area networks (WLAN) can cause an excessive level of interference in the commonly used 2.4 GHz ISM band, which leads to network performance degradation. This paper analyses the possibility to reduce congestion in the crowded ISM band by allowing some access points (APs) to opportunistically operate in a primary band. Although the problem of opportunistic channel allocation considering both ISM and primary bands can be formulated mathematically as a Binary Linear Programming (BLP) problem, existing algorithms (e.g., branch and bound) capable to provide optimal solutions for BLP problems are quite inefficient for large deployments. Therefore, we propose three different heuristic algorithms to solve this opportunistic channel allocation problem by properly extending graph coloring algorithms: First Fit (FF), Saturation Degree Ordering (SDO) and Minimum Spanning Tree (MST). A performance analysis of the three algorithms is carried out under different conditions of primary band availability and AP density