Vision-based plant growth monitoring system for tomatoes Brendick B. Abella...[et. al]

Thesis (Bachelor of Science in Electronics Engineering) -- University of Rizal System-Morong.

EXECUTIVE SUMMARY: The study aimed to achieve the following objectives: monitor the growth of tomato from transplanting stage to fruiting and detect parameters such as healthy and withered leaves, blossoming flowers and ripe and unripe fruits; determine the efficiency of the system through inspecting the tomato status with human vision and comparing the results with the results from the system; and determine the level of acceptability of the Vision-Based Plant Growth Monitoring System for Tomatoes in terms of applicability, completeness, functionality, usefulness and user-friendliness. The study was conducted at the University of Rizal System-Morong during the second semester of Academic year 2016-2017 up to the first semester of Academic year 2017-2018. The researchers made use of the experimental and descriptive method of research design. Qualitative description was used in the study and weighted mean used to determine the level of acceptability of the system in terms of applicability, user-friendliness, completeness, functionality and usefulness. The researchers used Mocrosoft Visual Studio 2013, C# as the programming language and integrated Aforge.NET library in developing the system. The level of acceptability in terms of applicability, user-friendliness, completeness, functionality and usefulness was determined through a questionnaire-checklist evaluated by a total of 40 respondents. They were divded into three groups, which consist of twenty (20) students, ten (10) professors and ten (10) farmers. The system was able to monitor and observe the growth of tomato plants parameters such as leaves, flowers and fruits. For the green leaves detection, the system detected 21 out of 20 green leaves detection, the system detected 21 out of 30 green leaves with a percentage of 70% while in terms of withered leaves detection, the system detected 1 out of 5 withered leaves with a percentage of 20%. For the flower detection, the system detected 12 out of 15 green leaves with a percentage of 80% while the system exhibited 100% accuracy in no flower detection. For the fruit detection, the system detected 8 out of 10 unripe fruits with a percentage of 80%, while the system detected 2 out of 5 ripe fruits with a percentage of 40%. The System exhibited 84% accuracy in no fruit detection. The system was also very much accepted as perceived by three groups of respondents. For its acceptability in terms of applicability, it was revealed that the system was very much accepted as evaluated by the students, professors and farmers since it gained an overall grand mean of 4.52. in terms of its user-friendliness, the system was very much accepted as evaluated by the students, professors and farmers as it gained an overall grand mean of 4.44. the system's acceptability in terms of completeness was also very much accepted as evaluated by the students, professors and farmers since it gained an overall grand mean of 4.37. in terms of its functionality, it was revealed that the system was very much accepted as evaluated by the students, professors and farmers since it gained an overall mean of 4.49. The following recommendations were draw based on the comments of the respondents as well as the researchers. The detection accuracy of the parameters must be improved, especially on the ripe fruits detection and withered leaves detection and add an additional feature of counting the number of fruits. It was recommended that in one click, all the parameters must be detected and displayed. The icons on the homepage of the program interface should also be labeled. Password security must be incorporated in the system. The system should be monitoring in real-time and the data should be sent remotely or uploaded in a website. Researchers may use different software applications and apply more deep-learning algorithm for the image processing.