In this paper we present a complete study on the balance between high performance image processing and low power consumption without using expensive components. Our proposal consists in implementing a Discrete Time Cellular Neural Network (DT-CNN) on a low power Actel IGLOO nano Field Programmable Gate Array (FPGA). This is a definitive step further from previous work to obtain an intelligent camera device for robots. Applications in Robot Guidance have rapidly increased in the last years as robots break in different fields of everyday live, which most of this robotic devices need sensors for navigation. Our proposed low cost solution avoids highly complex architectures, expensive smart sensors and low performance navigation systems.