Detection of cutting tool broken in CNC machining centers by image processing method

This thesis aims to determine if the drill or tap is broken as a result of the process performed with the drill and tap cutting tools used in CNC machining centers, with the camera system installed in the CNC machining center tool magazine and the machine integrated image processing software. First of all, the usage areas of image processing methods in the industry were examined in the thesis study. In the thesis study, image processing techniques were used with the MATLAB program. Raspberry Pi 3 and Raspberry Camera V2 module were used as camera and processor. The used machine in thesis is a Dahlih brand MCH-800 model horizontal machining center with Fanuc Oi-model D operating system. In the system established in practice, the necessary signaling structure has been created for the Raspberry Pi processor to work integrated with the CNC horizontal machining center. In order to make sense of the signals input/output from the processor in the PLC of the Fanuc Operating system, software has been added to the Ladder diagram according to the created algorithms. In this way, the Raspberry Pi camera module senses when to take a photo and which number drills or taps it takes. Drill and tap photos taken were saved in target files according to the software written in MATLAB program and classified according to image processing techniques and attributes. It has been determined that the drill or tap is broken because the value of the attributes changes when the drill or the tap with the features determined by the software is broken. This detection was given as information from the Raspberry Pi processor to the CNC machine and the CNC machine was given an alarm. As a result, the system designed and put into practice in practice has determined the fractures in drill and tap tools used in CNC machining centers by image processing method. With this thesis project, the applicability of image processing methods in the machining industry was supported.