<![CDATA[Currently, the car is one of the means of transportation that is widely used by many people and it has become a necessity to have a car to help users move more easily. Car technology continues to be developed by experts, including steering aid systems and safety for car users, such as automatic reading of objects and road boundaries that can be useful for both things. This system was built using the Fully Convolutional Network (FCN) method with Residual Neural Network (ResNet) architecture, and also Image Processing as signal processing with image input, and with a thermal Flir camera as vehicle input data. The data generated by this thermal camera is labeled first and then trained so that it can segment objects correctly according to their classification. In this study, the extraction accuracy of the training generated by the autonomous vehicle feature can reach 96.27% for ResNet 101 with a resolution of 640x480 pixels. As for suggestions for development to be even better in terms of segmentation, namely by using more training data than is used now and shooting locations for datasets in different places from the current research.]]>
Copyrights © 2022
