Identification of hopper gate sprocket during grain car unloading using digital image processing.
Mohan, A.L., Karunakaran, C., Jayas, D.S., and White, N.D.G. (2010). "Identification of hopper gate sprocket during grain car unloading using digital image processing.", Transactions of the ASABE, 53(4), pp. 1313-1320.
For most applications, machine vision solutions based on pattern recognition are developed using images acquired in a laboratory setting. Major constraints with these solutions occur when implementing them in real‐world applications. For instance, constantly changing ambient light conditions can pose many challenges to pattern recognition. The long‐term objective of this study is to automate the unloading of railroad grain cars in grain elevators. The first step for this automation is to correctly identify the hopper gate sprocket on the grain car. Algorithms were developed to detect and identify the sprocket under proper lighting conditions with 100% accuracy. The performance of the algorithms was also evaluated for the identification of the sprocket on a grain car exposed to different lighting conditions, which is expected to occur in typical grain unloading facilities. Monochrome images of the sprocket from a model system were acquired using different light sources, such as incandescent (direct or diffuse), fluorescent, and LEDs, with potential variables such as human presence behind the sprocket, stray light, or different backgrounds. To identify the sprocket, correlation and pattern recognition techniques using a template image combined with shape detection were used. The images were preprocessed using image‐processing techniques, including noise filtering and edge detection, prior to template matching. The template image developed from the light source that was similar to the light source used to acquire images simulating the work environment was more successful in identifying the sprocket. The template image used in the correlation technique easily identified the sprocket in the images from all light sources when none of the variables mentioned above were introduced. A combination of correlation with shape detection performed better than correlation alone in identifying the sprocket in images with external variables present.
Report a problem on this page
- Date modified: