Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/543023
Title: | Detection and Analysis System for Chemically Ripened Fruits Using Computer Vision |
Researcher: | Laxmi V |
Guide(s): | Roopalakshmi R |
Keywords: | Computer Science Computer Science Software Engineering Engineering and Technology |
University: | Visvesvaraya Technological University, Belagavi |
Completed Date: | 2023 |
Abstract: | Fruits have abundant source of vitamins, minerals, and fiber which are essential in fulfilling significant supplements of our day-to-day diet. However, nowadays, some greedy vendors use chemical ripening agents to ripen the fruits and consuming such fruits causes ill effects on the health of consumers. Though lot of efforts are made towards identification of chemically ripened fruits using computer vision yet, very less efforts are focused on detection of artificially matured mango fruits using hybrid features. To solve this issue this research work proposes three methodologies for detection, consumability analysis and prediction of chemically ripened fruits. newlineFirst methodology of the research work presents a GUI based identification of chemically ripened mango fruit using MPEG-7 descriptors. Specifically, in this framework visual based color features are used to detect chemically ripened mango fruits. In addition to this a robust detection framework for artificially and naturally ripened mangoes using SURF features is developed. In the second methodology, a consumability analysis using machine learning techniques is carried out. More specifically, this framework is proposed to analyze if a fruit is good to consume or not. Finally, in the third methodology, a CNN based prediction system which uses hybrid features is developed for predicting the nature of fruit. newlineThe three proposed methodologies are evaluated on the real time dataset which consists of app.5000 images of four different types of mangoes. Evaluations are carried out on the dataset and indicated in terms of metrics such as confusion matrix, precision and specificity. The performance results demonstrate the efficiency and effectiveness of the proposed framework in terms of time complexity, precision scores and confidence. newlineKeywords: Artificial Ripening Detection, Computer Vision, Deep Learning, Dominant colors, Speeded- Up Robust Features (SURF), Convolution Neural Networks. newlineiv newline |
Pagination: | 93 |
URI: | http://hdl.handle.net/10603/543023 |
Appears in Departments: | Department of Computer Science and Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_title.pdf | Attached File | 99.02 kB | Adobe PDF | View/Open |
02_prelim pages.pdf | 702.76 kB | Adobe PDF | View/Open | |
03_content.pdf | 74.98 kB | Adobe PDF | View/Open | |
04_abstract.pdf | 78.46 kB | Adobe PDF | View/Open | |
05_chapter 1.pdf | 403.85 kB | Adobe PDF | View/Open | |
06_chapter 2.pdf | 180.94 kB | Adobe PDF | View/Open | |
07_chapter 3.pdf | 1.16 MB | Adobe PDF | View/Open | |
08_chapter 4.pdf | 250.91 kB | Adobe PDF | View/Open | |
09_chapter 5.pdf | 294.93 kB | Adobe PDF | View/Open | |
10_annexures.pdf | 171.8 kB | Adobe PDF | View/Open | |
80_recommendation.pdf | 158.2 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: