Due to the various shapes, textures, and colors of fires, forest fire detection is a challenging task. The traditional image processing method relies heavily on manmade features, which is not universally applicable to all forest scenarios. In order to solve this problem, the deep learning technology is applied to learn and extract features of forest fires adaptively. However, the limited learning and perception ability of individual learners is not sufficient to make them perform well in complex tasks. Furthermore, learners tend to focus too much on local information, namely ground truth, but ignore global information, which may lead to false positives. In this paper, a novel ensemble learning method is proposed to detect forest fires in different scenarios. Firstly, two individual learners Yolov5 and EfficientDet are integrated to accomplish fire detection process. Secondly, another individual learner EfficientNet is responsible for learning global information to avoid false positives. Finally, detection results are made based on the decisions of three learners. Experiments on our dataset show that the proposed method improves detection performance by 2.5% to 10.9%, and decreases false positives by 51.3%, without any extra latency.

Forest fires have the characteristics of strong unpredictability and extreme destruction. Hence, it is difficult to carry out effective prevention and control. Once the fire spreads, devastating damage will be caused to natural resources and the ecological environment. In order to detect early forest fires in real-time and provide firefighting assistance, we propose a vision-based detection and spatial localization scheme and develop a system carried on the unmanned aerial vehicle (UAV) with an OAK-D camera. During the high incidence of forest fires, UAVs equipped with our system are deployed to patrol the forest. Our scheme includes two key aspects. First, the lightweight model, NanoDet, is applied as a detector to identify and locate fires in the vision field. Techniques such as the cosine learning rate strategy and data augmentations are employed to further enhance mean average precision (mAP). After capturing 2D images with fires from the detector, the binocular stereo vision is applied to calculate the depth map, where the HSV-Mask filter and non-zero mean method are proposed to eliminate the interference values when calculating the depth of the fire area. Second, to get the latitude, longitude, and altitude (LLA) coordinates of the fire area, coordinate frame conversion is used along with data from the GPS module and inertial measurement unit (IMU) module. As a result, we experiment with simulated fire in a forest area to test the effectiveness of this system. The results show that 89.34% of the suspicious frames with flame targets are detected and the localization error of latitude and longitude is in the order of 10⁻⁵ degrees; this demonstrates that the system meets our precision requirements and is sufficient for forest fire inspection.

Deep learning-based forest fire vision monitoring methods have developed rapidly and are becoming mainstream. The existing methods, however, are based on enormous amounts of data, and have issues with weak feature extraction, poor small target recognition and many missed and false detections in complex forest scenes. In order to solve these problems, we proposed a multi-task learning-based forest fire detection model (MTL-FFDet), which contains three tasks (the detection task, the segmentation task and the classification task) and shares the feature extraction module. In addition, to improve detection accuracy and decrease missed and false detections, we proposed the joint multi-task non-maximum suppression (NMS) processing algorithm that fully utilizes the advantages of each task. Furthermore, considering the objective fact that divided flame targets in an image are still flame targets, our proposed data augmentation strategy of a diagonal swap of random origin is a good remedy for the poor detection effect caused by small fire targets. Experiments showed that our model outperforms YOLOv5-s in terms of $mAP$ (mean average precision) by 3.2%, $A{P}_S$ (average precision for small objects) by 4.8%, $A{R}_S$ (average recall for small objects) by 4.0%, and other metrics by 1% to 2%. Finally, the visualization analysis showed that our multi-task model can focus on the target region better than the single-task model during feature extraction, with superior extraction ability.

This paper aims to solve multiple problems associated with attitude estimation; the complexity of Kalman Filter calculations in the Attitude and Heading Reference System (AHRS), the interference sensitivity of Magnetic Angular Rate and Gravity (MARG) sensors, and the low accuracy of the Factored Quaternion Algorithm (FQA). It presents a two-layer Linear Kalman Filter using MARG sensors to obtain attitude estimation in quaternions. Firstly, data from a triaxial accelerometer and magnetometer is processed by a novel algorithm, which fuses the Quaternion Estimator (QUEST) algorithm and FQA by the Linear interpolation (LERP) to obtain an observation model. Secondly, the process model in the two-layer Kalman Filter was obtained by using LERP to fuse an optimum quaternion obtained from a gyroscope and FQA. The LERP can eliminate gyro bias drift, and integral error, and compensate for unexpected conditions, such as fast rotation and temporary strong magnetic disturbances. The proposed algorithm presents higher accuracy and lower computational load than QUEST or FQA used with a Kalman Filter alone. The performance of the proposed algorithm is verified through simulation and experiments.

The inertial navigation system and celestial navigation system (CNS) are important components of autonomous satellite navigation, but each has inherent limitations. INS relies on accelerometers and gyroscopes, whose bias errors, particularly gyro bias, accumulate over time and cause significant navigation drift. Conventional CNS uses a star tracker to provide attitude information and combines it with a horizon sensor or starlight refraction measurements to determine position. However, it cannot directly obtain positional information from star trackers and relies on indirect calculations based on atmospheric starlight refraction. This approach is limited to a refraction height range of approximately 20–50 km and requires time-consuming star-image simulation and matching.To address these limitations, this paper proposes an autonomous navigation system integrating an IMU with two star trackers. One star tracker observes through the Earth’s atmosphere, while the other provides unobstructed stellar measurements and is used to correct IMU errors, thereby mitigating gyro bias. The two star trackers are mounted at a fixed relative angle, allowing the use of their measured angular relationship instead of computationally expensive star-image simulation and attitude transformation.The proposed system operates in two modes. When the starlight refraction height is between 20 and 50 km, a single CNS provides navigation. Outside this range, INS and CNS are integrated using a federated Unscented Kalman Filter (FUKF). The proposed system achieves meter-level position accuracy and improved velocity estimation. Its performance is validated using flight data from the CUAVA-2 CubeSat, demonstrating its potential for robust autonomous satellite navigation.