Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/302188
Title: | Jasclassify hybrid algorithm for quality detection in Jasmenum flowers |
Researcher: | Krishnaveni S. |
Guide(s): | Pethalakshmi A. |
Keywords: | Computer Science Computer Science Artificial Intelligence Engineering and Technology |
University: | Mother Teresa Womens University |
Completed Date: | 2019 |
Abstract: | Jasmine floriculture is one of the leading flower framing in India, as it produces more returns to the smalls farmers. Across the country, Andra Pradesh and Tamil Nadu are the two major states has higher percentage of jasmine flower production. Particularly in Tamil Nadu, the Dindigul district contributes 14% of overall production, ranked first in state. Jasmine flower has good market in export, oil and perfume industries. One major issue in jasmine farming is that the labor cost contributes 28% among overall establishment cost, this would increase as much as manpower involved. Also, due to manual intervention, the marketing process might get delayed, at the same time the jasmine flower may lost is freshness, which affects the returns too. This research work focused to reduce the labor time involved in process of partitioning the jasmine flowers into normal and defected based on their freshness and quality. The automated jasmine flower classification (JasClassify) algorithm makes use of image processing methods to classify the flowers into normal and defected. Initially the jasmine flower images are captured through a digital camera. In the second step the acquired image is segmented to partition the jasmine flower region while discarding the background. A novel MultiHistogram based Otsu Thresholding (MHOT) method is proposed for segmentation. In the third step, the color, texture and shape feature descriptors are extracted from the segmented image. Average Color Differences (ACD), Color Edge Directivity Descriptor (CEDD), Local Binary Patterns (LBP) and Zernike Moment (ZM) methods are applied to extract the feature descriptors. These feature vectors are normalized and fused to form one single feature vector for about 500 images in the dataset. newline |
Pagination: | 245p. |
URI: | http://hdl.handle.net/10603/302188 |
Appears in Departments: | Department of Computer Science |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_title.pdf | Attached File | 2.44 MB | Adobe PDF | View/Open |
02_certificate.pdf | 260.17 kB | Adobe PDF | View/Open | |
03_abstract.pdf | 98.21 kB | Adobe PDF | View/Open | |
04_declaration.pdf | 244.93 kB | Adobe PDF | View/Open | |
05_plagiarism certificate.pdf | 215.66 kB | Adobe PDF | View/Open | |
06_acknowledgement.pdf | 177.51 kB | Adobe PDF | View/Open | |
07_contents.pdf | 111.84 kB | Adobe PDF | View/Open | |
08_list of tables.pdf | 191.33 kB | Adobe PDF | View/Open | |
09_list of figures.pdf | 2.45 MB | Adobe PDF | View/Open | |
10_abbreviation.pdf | 2.62 MB | Adobe PDF | View/Open | |
11_chapter1.pdf | 3.04 MB | Adobe PDF | View/Open | |
12_chapter2.pdf | 2.74 MB | Adobe PDF | View/Open | |
13_chapter3.pdf | 3 MB | Adobe PDF | View/Open | |
14_chapter4.pdf | 3 MB | Adobe PDF | View/Open | |
15_chapter5.pdf | 3.06 MB | Adobe PDF | View/Open | |
16_chapter6.pdf | 2.4 MB | Adobe PDF | View/Open | |
17_conclusion.pdf | 2.45 MB | Adobe PDF | View/Open | |
18_bibiliography.pdf | 2.68 MB | Adobe PDF | View/Open | |
80_recommendation.pdf | 229.84 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: