Unmanned Aerial Vehicles (UAVs) are becoming a significant field of research with numerous applications, ranging from mapping to surveillance. New applications, such as aerial delivery of goods, are expected to appear in the next years and will require more and more autonomy from UAVs. One challenge preventing UAVs from being fully autonomous is their current limitations in handling potential collisions among multiple vehicles. This paper presents a collision avoidance algorithm for fixed-wing UAVs navigating in a three dimensional space. It satisfies limited field of view constraints that stem from the use of a single camera system as sensing device. The proposed algorithm uses potential fields to both navigate and avoid obstacles. To guarantee collision avoidance, the algorithm is enhanced with a turning behavior that allows for ensuring the safety of the method. Simulations are performed to show the effectiveness of the proposed algorithm.