Development Of Artificial Intelligence (Ai) For Image Processing And Conceptual Distance Computation From Camera For Pick And Place Of Oil Palm Fresh Fruit Bunch (Ffb)
In Malaysia, agriculture activities are the major sector that provides importance to the entire human beings especially in the food industry. Most of the researchers have done research in agriculture every day to aim for development in quality, productivity and limiting probability of human er...
Saved in:
| Main Author: | |
|---|---|
| Format: | Monograph |
| Language: | English |
| Published: |
Universiti Sains Malaysia
2019
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/58302/1/Development%20Of%20Artificial%20Intelligence%20%28Ai%29%20For%20Image%20Processing%20And%20Conceptual%20Distance%20Computation%20From%20Camera%20For%20Pick%20And%20Place%20Of%20Oil%20Palm%20Fresh%20Fruit%20Bunch%20%28Ffb%29.pdf http://eprints.usm.my/58302/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In Malaysia, agriculture activities are the major sector that provides importance to
the entire human beings especially in the food industry. Most of the researchers have done
research in agriculture every day to aim for development in quality, productivity and
limiting probability of human error. Malaysia is categorized as the largest producer and
exporter of oil palm globally. In order to increase the productivity of palm oil fruits, various
machine-vision techniques and mechanization systems can be applied. This is to reduce the
human labor of picking the crops by simply using sickles. The labor shortage problems can
be minimized as not many labors are willing to do the hard job while the salary is not
reasonable. Therefore, the idea of this project is to create a system for a machine to pick and
place the crops using an automated bot mounted at the truck with the aid of live feed camera.
As early steps, a system of oil palm FFB detection is being created using two algorithms of
image processing, YOLO and Tensorflow Object Detection API. The performance of both
algorithms are then determined in terms of confidence level and speed to process the image
in three different parameters. Then, a study of how the camera computes the distance of
detected oil palm FFB from the camera is done in this project. Throughout this project, it
can be deduced that YOLO provides higher speed while Tensorflow performs better
accuracy in detecting the oil palm FFB. A concept to compute the distance between camera
and the respective oil palm fruit is proposed in this research for future work application and
experiment. |
|---|