CUR - The Council on Undergraduate Research

Application to Analyze Color Calibration Needed for Automated Grinding of Metalcastings

Aaron Bodenham and Dr. Frank Peters, Department of Industrial and Manufacturing Systems Engineering, Iowa State University, Riggs Ct, Ames, IA 50010.

Manual grinding of a metalcasting is a difficult task, which motivated the development of an automated grinding system for use in foundries. Automation of this process would improve work conditions within the foundry. The automated grinding method utilizes markings on castings drawn by operators to automatically create a path plan for material removal. A 3D scanner mounted to the robot records an XYZ point cloud of the casting as well as the color information. The color information will be used to isolate the markings made by the operator, which will identify the areas to be ground and used to segment the point cloud. Color thresholds must be established accurately in order for the automated system to function. Due to light conditions, surface finish, and marker color, not all operator marks will be detected through the scanner without setting proper thresholds. An application was developed using MATLAB that will be used by foundries to assist in setting up automated grinding systems, starting with a scan of the casting and its environment. The user will begin with picking three points from the cloud and recording the RGB values. An average of these values will be used, then have a dynamic increment and decrement applied to each color section, based on the range of values within the points selected. This app allows users to use either RGB or HSV color spaces, depending on preference. Testing will be conducted using this app, providing an initial recommendation to foundries for optimal color choice. Once the optimal color is determined for marking anomalies, the grinding program can be used to accurately interpret colored marks on the casting and begin removing material.

Additional Abstract Information