In recent years, outdoor mobile augmented reality has garnered significant attention due to its potential applications in navigation, tourism, education, emergency and disaster management, the environmental sector, and entertainment. However, the user experience and effectiveness of these applications are heavily reliant on precise geo-localization, a challenge that remains largely unsolved in rural, non-mountainous outdoor environments. This paper aims to contribute to the field of geo-localization for outdoor mobile augmented reality by providing a comprehensive overview of state-of-the-art visual geo-localization methods, introducing a novel classification, and emphasizing the demand and research gap in this field. An analysis of existing methods reveals their limited applicability in rural, non-mountainous outdoor environments, as they are often reliant on either street view imagery or distinct mountain silhouettes and lack the capability of performing 6D localization, which is crucial for augmented reality applications. The analysis further emphasizes the limitations of visual geo-localization datasets for outdoor mobile augmented reality, underscoring the necessity for new datasets to address these limitations.