<![CDATA[In this final project, a spherical object detection system has been designed in which the final result will display the class name according to the detected object and a bounding box on the object indicating the object is detected accordingly. What will be done from having a dataset of 202 images and divided into 70% training data, 20% validation data, 10% test data. By using the YOLOv4 method, it is hoped that the detection of spherical objects will be more efficient in detecting an object that is needed, the final result of the implementation of this spherical object detection system will display a bounding box and the accuracy of objects detected on the laptop screen for testing and analysis results of the YOLOv4 method performance system. done by confusion matrix which calculates the results of accuracy, recall, precision and there are several tests to find out with different conditions the system can detect an object. In the first test, the ball was detected by being blocked by another object in the percentage value of 50%, 60%, 70% of the system being able to detect a ball object that was blocked by another object, then with an obstacle value of 80%, 90%, 100% the system could not detect a ball object.]]>
Copyrights © 2022
