Swarm robots are a large group of simple robots that coordinate their actions through decentralization. Individual robots make decisions based on interactions and information gathered from the environment. This enables the bot to mimic the natural swarm behaviour seen in birds, ants, and other natural insects. This research proposes using Computer Vision to intervene if a bot fails to exploit adaptive environmental variables to adjust a decentralized swarm’s autonomous behaviour. The decision of selecting a computational resource for processing of real-time detection and tracking of bio-inspired Kilobot is challenging. A comparison between local and cloud-based computational resources has been presented utilising varying environmental conditions which can help determine the most efficient and affordable platform for swarm processing. Local computational resources provide overall efficient results but Cloud computational resources provide faster execution results.