This document discusses collision avoidance in CNC machines using multiple sensors. It proposes detecting collisions between the cutting tool and workpiece using a Kinect sensor and two webcams through image processing techniques in MATLAB. The methodology involves extracting depth maps from Kinect to detect the spindle and workpiece, and using SIFT and normalized cross-correlation on webcam images to detect the cutting tool. Adaptive background subtraction will be used to isolate foreground objects. Detected objects across sensors will be fused to determine if a collision has occurred.