Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Machine vision solutions based on pattern recognition techniques for most applications are developed using images acquired in a laboratory setting. One of the major constraints with these solutions incurs when implementing them for real-time applications. For instance, constant changing ambient lighting conditions may pose many challenges to pattern recognition. The long term objective of this study is to automate the unloading of hopper grain-cars in grain elevators. In this manuscript, identification of the sprocket in a grain-car, exposed to different lighting conditions simulating the real-time environment, is described. To identify the sprocket, correlation, a pattern recognition technique and shape detection were used. The images were preprocessed using image processing techniques including filtering and edge detection prior to template matching. Templates developed from light sources which were similar to lights used in the work environments were more successful in identifying the sprocket. A combination of correlation and shape detection algorithm performed better than correlation alone in identifying the sprocket.