This work presents a method for urban aerial route planning ("drone corridors") in a real-world environment using single-solution metaheuristics. The central region of Belo Horizonte was modeled as a geospatial graph with regulatory constraints and a detailed risk map based on population density, buildings, and critical infrastructure. Routes are optimized using Simulated Annealing (SA) and Tabu Search (TS), considering distance, risk, and time objectives. Six operational scenarios are evaluated. The results show that both techniques generate feasible and safe routes, with relevant differences in exploration and convergence behavior. The proposed approach demonstrates practical applicability for supporting UAV Traffic Management (UTM) systems in dense urban environments.