• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.5
• /
• pp.19-24
• /
• 2024

Global climate change is causing an increase in natural disasters around the world due to long-term temperature increases and changes in rainfall. Among them, forest fires are becoming increasingly large. South Korea experienced an average of 537 forest fires over a 10-year period (2013-2022), burning 3,560 hectares of forest. That's 1,180 soccer fields(approximately 3 hectares) of forest burning every year. This paper proposed an artificial intelligence based wildfire detection method using image and text information. The performance of the proposed method was compared with YOLOv9-C, RT-DETR-Res50, RT-DETR-L, and YOLO-World-S methods for mAP50, mAP75, and FPS, and it was confirmed that the proposed method has higher performance than other methods. The proposed method was demonstrated as a forest fire detection model of the early forest fire detection system in the Gangwon State, and it is planned to be advanced in the direction of fire detection that can include not only forest areas but also urban areas in the future.

• Korean Journal of Artificial Intelligence
• /
• v.12 no.4
• /
• pp.27-33
• /
• 2024

Various techniques are being researched to effectively detect forest fires. Among them, techniques using object detection models can monitor forest fires over wide areas 24 hours a day. However, detecting forest fires early with traditional object detection models is a very challenging task. While they show decent accuracy for thick smoke and large fires, they show low accuracy for faint smoke and small fires, and frequently generate false positives for lights that are like fires. In this paper, to solve these problems, we focus on leveraging local characteristics such as contours and textures of fire and smoke, which are crucial for accurate detection. Based on this approach, we propose EDAM (Edge driven Attention Module) that performs enhancement by richly utilizing contour and texture information of fire and smoke. EDAM extracts important edge information to generate feature maps with emphasized contour and texture information, and based on this map, performs Attention Mechanism to emphasize key characteristics of smoke and fire. Through this mechanism, the overall model performance was improved, with AP_sincreasing from 0.154 to 0.204 and AP_0.5 from 0.779 to 0.784, resulting in a significant improvement in AP_svalue to 32.47%. In practice, the model applying this technique showed excellent inference speed while greatly improving detection performance for small objects compared to existing models and reduced false positive rates for building and street light illumination in nighttime environments that are easily mistaken for fire.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.54-57
• /
• 2022

Korea has a vulnerable environment from the risk of wildfires, which causes great damage every year. To prevent this, a lot of manpower is being used, but the effect is insufficient. If wildfires are detected and extinguished early through artificial intelligence technology, damage to property and people can be prevented. In this paper, we studied the object double detection method with the goal of minimizing the data collection and processing process that occurs in the process of creating an object detection model to minimize the damage of wildfires. In YOLOv5, the original image is primarily detected through a single model trained on a limited image, and the object detected in the original image is cropped through Crop. The possibility of improving the false positive object detection rate was confirmed through the object double detection method that re-detects the cropped image.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.58 no.2
• /
• pp.142-148
• /
• 2013

Bacterial diseases in soybean are bacterial pustule by Xanthomonas axonopodis pv. glycines, wildfire by Pseudomonas syringae pv. tabaci, bacterial blight by Pseudomonas savastanoi pv. glycines and bacterial brown spot by Pseudomonas syringae pv. syringae in Korea. It is difficult to identify each disease by early symptoms in fields, because the initial symptoms of these diseases are very similar to each other. In this study, we developed multiplex PCR detection method for rapid and accurate diagnosis of bacterial diseases. The glycinecin A of X. axonopodis pv. glycines, the tabtoxin of P. syringae pv. tabaci, the coronatine of P. savastanoi pv. glycines and the syringopeptin of P. syringae pv. syringae have been reported previously. These bacteriocin or phytotoxin producing genes were targeted to design the specific diagnostic primers. The primer pairs for diagnosis of each bacterial diseases were selected without nonspecific reactions. The studies on simultaneous diagnosis method were also conducted with primarily selected 21 primers. As a result, we selected PCR primer sets for multiplex PCR. Sizes of the amplified PCR products using the multiplex PCR primer set consist of 280, 355, 563 and 815 bp, respectively. This multiplex PCR method provides a efficient, sensitive and rapid tool for the diagnosis of the bacterial diseases in soybean.