A unique multisensory tool monitoring system using machine vision and vibration sensors has been developed for turning operations. Vibration signals from the tool are monitored on-line using a piezoelectric accelerometer mounted on the tool holder and the tool condition is monitored periodically using a vision sensory system. Images of the tool are frame grabbed between cuts. The developed image processing algorithm uses neural network mathematics to measure the tool wear in the captured images. The inputs to the neural network are a set of vectors defining the characteristics of the tool surface such as surface curvature. Features such as area, perimeter, width and height of the worn area are measured to identify the extent of the tool wear. The system is unique in that it combines an indirect tool monitoring technique, vibration monitoring, with a direct visual monitoring technique. The addition of the vision system increases the reliability of the monitoring system by detecting false signals received from the vibration sensor. The vibration sensor, on the other hand, monitors the tool condition on-line and can detect tool breakage. By using both vibration and vision sensory a more accurate tool monitoring system is developed; false calls that may result if any one of these systems is used independently are reduced. © 1997 Taylor & Francis Ltd.