Forest fires are one of the most important threats for forests in the State of Mexico. Therefore, understanding their geographical patterns is a priority for the design of forest management strategies. We processed th...Forest fires are one of the most important threats for forests in the State of Mexico. Therefore, understanding their geographical patterns is a priority for the design of forest management strategies. We processed the records obtained with the MOD14A2 product (for thermal anomalies and fire) of MODIS sensor. Such scenes correspond to dry seasons (from March 15 to June 30) from 2000 to 2012 in the State of Mexico. We analyzed such records in a GIS environment to learn their spatial patterns and establish their geographical correlations as a first step to understand the causal agents of forest fires. As a result, forest fires in the State of Mexico showed a clustered spatial trend with a southwest tendency and a slight spatial relation with total winter precipitation and maximal temperature in summer.展开更多
To improve the performance of the forest fire smoke detection model and achieve a better balance between detection accuracy and speed, an improved YOLOv4 detection model (MoAm-YOLOv4) that combines a lightweight netwo...To improve the performance of the forest fire smoke detection model and achieve a better balance between detection accuracy and speed, an improved YOLOv4 detection model (MoAm-YOLOv4) that combines a lightweight network and attention mechanism was proposed. Based on the YOLOv4 algorithm, the backbone network CSPDarknet53 was replaced with a lightweight network MobilenetV1 to reduce the model’s size. An attention mechanism was added to the three channels before the output to increase its ability to extract forest fire smoke effectively. The algorithm used the K-means clustering algorithm to cluster the smoke dataset, and obtained candidate frames that were close to the smoke images;the dataset was expanded to 2000 images by the random flip expansion method to avoid overfitting in training. The experimental results show that the improved YOLOv4 algorithm has excellent detection effect. Its mAP can reach 93.45%, precision can get 93.28%, and the model size is only 45.58 MB. Compared with YOLOv4 algorithm, MoAm-YOLOv4 improves the accuracy by 1.3% and reduces the model size by 80% while sacrificing only 0.27% mAP, showing reasonable practicability.展开更多
文摘Forest fires are one of the most important threats for forests in the State of Mexico. Therefore, understanding their geographical patterns is a priority for the design of forest management strategies. We processed the records obtained with the MOD14A2 product (for thermal anomalies and fire) of MODIS sensor. Such scenes correspond to dry seasons (from March 15 to June 30) from 2000 to 2012 in the State of Mexico. We analyzed such records in a GIS environment to learn their spatial patterns and establish their geographical correlations as a first step to understand the causal agents of forest fires. As a result, forest fires in the State of Mexico showed a clustered spatial trend with a southwest tendency and a slight spatial relation with total winter precipitation and maximal temperature in summer.
文摘To improve the performance of the forest fire smoke detection model and achieve a better balance between detection accuracy and speed, an improved YOLOv4 detection model (MoAm-YOLOv4) that combines a lightweight network and attention mechanism was proposed. Based on the YOLOv4 algorithm, the backbone network CSPDarknet53 was replaced with a lightweight network MobilenetV1 to reduce the model’s size. An attention mechanism was added to the three channels before the output to increase its ability to extract forest fire smoke effectively. The algorithm used the K-means clustering algorithm to cluster the smoke dataset, and obtained candidate frames that were close to the smoke images;the dataset was expanded to 2000 images by the random flip expansion method to avoid overfitting in training. The experimental results show that the improved YOLOv4 algorithm has excellent detection effect. Its mAP can reach 93.45%, precision can get 93.28%, and the model size is only 45.58 MB. Compared with YOLOv4 algorithm, MoAm-YOLOv4 improves the accuracy by 1.3% and reduces the model size by 80% while sacrificing only 0.27% mAP, showing reasonable practicability.
