It is very important for us to rescue from fatal accidents caused by fire. Recently in Bangladesh more than two hundred garment workers have deceased at Tajrin Fashion Industry. In this work we have performed the edge detection for Robot on the eve of edge identification to save the worker while they will be locked at emergency situations where human interaction will be failed. We have trained the system such a way that a Robot can easily learn the situations. We have used Automated Brained Learning (ABL) for Robot to detect the objects. Our work ensures only the edge detection. Sobel operator and masking is used in this process. To accomplish the work RGB color model and other color model such as YMC color model is analyzed to ensure the better result. We have noticed that RGB color model is better for our ABL process. Besides, YMC color model also generate good result while the fire is over smoked.