Computer vision algorithms for on-tree fruit detection and counting of mango fruit images | |
| Author | Dodeja, Neha |
| Call Number | AIT Caps. Proj. no.TC-14-02 |
| Subject(s) | Computer vision Image processing Mango |
| Note | A project submitted in partial fulfillment of the requirements for the Degree of Bachelor of Science in Engineering in Telecommunications Engineering, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Caps. Proj. ; no. TC-14-02 |
| Abstract | The most ancient and yet economically crucial activity for mankind is agriculture which provides the essential fuel, food and fiber for the survival of humans. As population continues to grow worldwide with the expectation to breach the 9 billion mark by the fifth decade of the century, production in the farms needs to increase twofold if it must match the ever increasing bio-energy and food requirement. As our physical resources i.e. water, arable land and resources are limited; estimates suggest that production in agricultural farms must increase its efficiency by 25% so that this goal can be achieved. All of this must be achieved without any additional pressure on the environment. Automation of agricultural processes, especially harvesting, in the form of agricultural robots can be used to meet the food needs of the society by 2050. Hence, harvesting by automatic processes is currently a fast developing area of research. This research presents three different computer vision algorithms which can be used for automatic on-tree fruit detection and counting namely, template matching, segmentation and blob detection and eigenfruit detection. For this study, a database of night time images of mango trees was used. All the algorithms were implemented on MATLAB and the results were calibrated on the dataset. The results were then analyzed on the basis of number of fruits detected, the hit-rates and miss-rates. Performance measures were calculated. Template matching algorithm showed a precision rate of 82.02% and segmentation method showed a precision rate as high as 90%. Based on these results a regression model was also developed. While these techniques require further development, they have opened a wide area for application and expansion of computer vision in fruit harvesting. This is just the beginning of what will be a revolution in the way that crops are grown, harvested and tended. |
| Year | 2014 |
| Corresponding Series Added Entry | Asian Institute of Technology. Caps. Proj. ; no. TC-14-02 |
| Type | Capstone Project |
| School | School of Engineering and Technology (SET) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Telecommunications (TC) |
| Chairperson(s) | Attaphongse Taparugssanagorn; |
| Examination Committee(s) | Mongkol Ekpanyapong; |
| Degree | Capstone Project (B. Sc.) - Asian Institute of Technology, 2014 |