Early wildfire detection is crucial for minimizing environmental and societal impacts. This paper tackles the Camera Placement Optimization (CPO) problem for fire monitoring, optimizing camera placement and orientation to maximize risk-weighted coverage while accounting for visibility constraints, which is different from traditional approaches that prioritize total area coverage. Leveraging their advantages in scalability, exploration granularity and abstraction from the specific object function formulation, we employ three metaheuristic techniques, Genetic Algorithms (GA), Tabu Search (TS), and Particle Swarm Optimization (PSO), to efficiently explore the solution space. A large-scale case study in western Piedmont, Italy, demonstrates that all three methods can converge and outperform the baseline, particularly in densely packed scenarios where coverage zones overlap. While all three algorithms present comparable performance levels, GA and TS have a lead in simpler scenarios with fewer cameras, while PSO excels in more complex configurations.