Automatic Barley Relevance and Quality Assessment for Brewing Industry
| dc.contributor.advisor | Assabie, Yaregal (PhD) | |
| dc.contributor.author | Adugna, Mesfin | |
| dc.date.accessioned | 2019-04-22T07:21:50Z | |
| dc.date.accessioned | 2023-11-09T11:25:39Z | |
| dc.date.available | 2019-04-22T07:21:50Z | |
| dc.date.available | 2023-11-09T11:25:39Z | |
| dc.date.issued | 2017-11-02 | |
| dc.description.abstract | In beer production malt barley is used as major raw material holding about 90% of the total raw material cost. Homogeneity, a measure of grain uniformity, is important for malting and brewing performance. Uniformity of malt barley might be influenced due to different reasons that result in loss of quality during beer production, which has direct impact on brewing industry and on their customer satisfaction. Currently, malt barley sample quality inspection to keep homogeneity is performed manually by human experts through visual evaluation. However, visual evaluation requires significant amount of time, trained and experienced people. Besides, it is affected by bias and inconsistencies associated with human nature. Such approach will not be satisfactory for large scale inspection and grading unless fully automated. The goal of this research work is to assess the quality of malt barley sample by identifying objects found in the image and giving grade level category for malt barley grains using digital image processing techniques, based on the specification standard for malt barely by Ethiopian Standards Agency. The system architecture for assessing malt barley quality sample consists of four components namely preprocessing, segmentation, feature extraction, and classification. Preprocessing convert RGB image to grayscale, filter noise, and apply binarization. A new segmentation that effectively segments malt barley grains from the background is developed from combination of three existing edge detection methods. We also applied ellipse fitting model on contour estimation method to segment overlapping malt barley grains. A total of 17 (1 size, 7 shape, 3 texture, and 6 color) features are identified to model the objects found in malt barley sample. For the purpose of classification Artificial Neural Network is used, the training data is randomly partitioned into training (70%) and testing (30%). The classifier achieved an overall classification accuracy of 99.4 % for grading malt barley level and 98.7% for classification of object class. | en_US |
| dc.identifier.uri | http://10.90.10.223:4000/handle/123456789/18107 | |
| dc.language.iso | en | en_US |
| dc.publisher | Addis Ababa University | en_US |
| dc.subject | Malt Barley | en_US |
| dc.subject | Edge Segmentation | en_US |
| dc.subject | Contour Estimation | en_US |
| dc.subject | Overlapped Segmentation | en_US |
| dc.subject | Artificial Neural Network | en_US |
| dc.subject | Digital Image Processing | en_US |
| dc.subject | Homogeneity | en_US |
| dc.subject | Quality Assessment | en_US |
| dc.title | Automatic Barley Relevance and Quality Assessment for Brewing Industry | en_US |
| dc.type | Thesis | en_US |