<![CDATA[The land area of a country has a state border area that can occur threats and conflicts such as tank vehicles that can penetrate into the territory of another country. The threat is overcome by object detection technology to detect tanks in land surveillance using the YOLO model. This research aims to obtain the results of the tank detection system data parameters, determine the tank detection process, and determine the effectiveness of the YOLO model in detecting tanks. The YOLO model used in the research is YOLOv9 by collecting tank image datasets to serve as a training dataset that will produce data parameters in object detection and evaluating model testing in the form of images, videos, and detection devices in the form of embedded systems integrated with cameras. Evaluation of tank detection in the form of simulation is tested using three different confidence values and using a dark or night scenario to determine the effectiveness of the YOLOv9 model in detecting tanks in that scenario. The results of the evaluation of the YOLOv9 model in tank image detection get 99.3% accuracy in the form of images and videos. The evaluation results in the scenario of three different confidence values can sort out the low accuracy value of tank detection according to the confidence value used and the results of tank detection in dark or night conditions are less effective in detecting tanks. This research is concluded to be able to produce tank detection system data parameters in the form of precision and recall, can find out the tank detection process, and the YOLO model becomes an effective object detection model in detecting tanks from the detection simulation results obtained.]]>
