Forest fires are a huge ecological hazard, and smoke is an early characteristic of forest fires. Smoke present only in tiny region images that captured the stages occurrence or when far from camera. Furthermore, dispersal uneven, background environment complicated changing, thereby leading to inconspicuous pixel-based features complicate detection. In this paper, we propose detection method called multioriented based on value conversion-attention mechanism module Mixed-NMS (MVMNet). First,...

After the birth of deep learning, artificial intelligence has entered a vigorous period rapid development. In this process rising and growing, we have made one achievement after another. When learning is applied to fruit target detection, due complex recognition background, large similarity between models, serious texture interference, partial occlusion fruits, detection rate based on traditional methods low. order solve these problems, BCo-YOLOv5 network model proposed recognize detect...

Graph convolutional networks (GCNs) are a promising approach for addressing the necessity long-range information in hyperspectral image (HSI) classification. Researchers have attempted to develop classification methods that combine strong generalizations with effective However, current HSI based on GCN present two main challenges. First, they overlook multi-view features inherent HSIs, whereas interacts each other facilitate tasks. Second, many algorithms perform rudimentary fusion of...

The target detection of smoke through remote sensing images obtained by means unmanned aerial vehicles (UAVs) can be effective for monitoring early forest fires. However, targets in UAV are often small and difficult to detect accurately. In this paper, we use YOLOX-L as a baseline propose network based on the parallel spatial domain attention mechanism small-scale transformer feature pyramid (PDAM–STPNNet). First, enhance proportion fire dataset, component stitching data enhancement generate...

Grape disease is a significant contributory factor to the decline in grape yield, typically affecting leaves first. Efficient identification of leaf diseases remains critical unmet need. To mitigate background interference feature extraction and improve ability extract small spots, by combining characteristic features diseases, we developed novel method for recognition classification this study. First, Gaussian filters Sobel smooth de-noising Laplace operator (GSSL) was employed reduce image...

In deep learning-based maize leaf disease detection, a identification method called Network based on wavelet threshold-guided bilateral filtering, multi-channel ResNet, and attenuation factor (WG-MARNet) is proposed. This can solve the problems of noise, background interference, low detection accuracy images. To begin, processing layer Wavelet threshold guided filtering (WT-GBF) WG-MARNet model employed to reduce image noise perform high low-frequency decomposition input using WT-GBF....

The appearance quality of apples directly affects their price. To realize apple grading automatically, it is necessary to find an effective method for detecting surface defects. Aiming at the problem a low recognition rate in defect detection under small sample conditions, we designed network (ASDINet) suitable learning. self-developed sorting system collected RGB images 50 samples model verification, including non-defective and defective (rot, disease, lacerations, mechanical damage)....

Tomato is an important and fragile crop. During the course of its development, it frequently contaminated with bacteria or viruses. leaf diseases may be detected quickly accurately, resulting in increased productivity quality. Because intricate development environment tomatoes their inconspicuous disease spot features small area, present machine vision approaches fail to reliably recognize tomato leaves. As a result, this research proposes novel paradigm for detecting disease. The INLM...

Accurate segmentation of the stem pumpkin seedlings has a great influence on modernization cultivation, and can provide detailed data support for growth plants. We collected constructed seedling point cloud dataset first time. Potting soil wall background in often interfere with accuracy partial cutting stems. The shape varies due to other environmental factors during growing stage. is closely connected potting leaves, boundary easily blurred. These problems bring challenges accurate In this...

Affected by various environmental factors, citrus will frequently suffer from diseases during the growth process, which has brought huge obstacles to development of agriculture. This paper proposes a new method for identifying and classifying diseases. Firstly, this designs an image enhancement based on MSRCR algorithm homomorphic filtering optimized Laplacian (HFLF-MS) highlight disease characteristics citrus. Secondly, we designed neural network DS-MENet DenseNet-121 backbone structure. In...

The classification of ground objects from hyperspectral images (HSIs) is great importance for human perception information about the terrain and landscape. HSIs have numerous dimensions, obtaining data difficult. issue slow convergence neural network training brought on by high dimensional data, network's performance impacted challenging acquisition process. In order to achieve effects low dependence rapid convergence, we propose a redundancy elimination architecture with decoupled-gaze...

Health insurance is a type of that covers individual and family medical expenses important for the health financial security individuals families. To better predict demand insurance, three regression models in machine learning - random forest, linear regression, decision tree are widely used prediction. Among these models, forest has best prediction effect with model score 0.8564, which among models. Random an integrated method combines multiple into more powerful can effectively avoid...

Clusters of dead trees are forest fires-prone. To maintain ecological balance and realize its protection, timely detection in remote sensing images using existing computer vision methods is great significance. Remote captured by Unmanned aerial vehicles (UAVs) typically have several issues, e.g., mixed distribution adjacent but different tree classes, interference redundant information, high differences scales clusters, making the clusters much more challenging. Therefore, based on Multipath...

Bearings are crucial components of mechanical and electrical systems. However, they susceptible to damage, which significantly impact equipment performance safety. In this paper, we explore the effectiveness vibration current signals as technologies for condition monitoring bearings. To enhance accuracy fault detection, propose a novel approach that signal fused generate symmetrized dot pattern (SDP) image. Furthermore, parameters SDP images optimized using structural similarity index...

Accurate segmentation of the stem pumpkin seedlings has a great influence on modernization cultivation, which can provide detailed data support for growth plants. We collected and constructed seedling point cloud dataset first time. Potting soil wall background in often interfere with accuracy partial cutting stems. The shape varies due to other environmental factors during growing stage. is closely connected potting leaves, boundary easily blurred. These problems bring challenges accurate...

In this paper, we performed bitcoin price prediction based on dataset using Support Vector Machine model, Random Forest Neural Network XGBoost model and LightGBM evaluated the performance of these models. We divided Bitcoin into training test sets in a ratio 7:3, where 70 were used as set 30 set. The models trained with tested stock change (yield) target variable other variables input variables. By comparing MSE, RMSE, MAE, MAPE R different it was found that has best prediction. four ranged...

This paper proposes a novel multi-coil inductive displacement measurement sensor that can achieve high-precision position calculation. Conventionally, precise adjustment of windings is made to enhance the linearity range sensor. For proposed in this paper, finite element method was used analyze dynamic characteristics The results shows there linear region between induced voltage sensing coils and core. A high precision calculation paper. solves problem nonlinear affecting accuracy...

Multi target tracking refers to the use of data obtained by radar sensors track and identify position, speed other related attributes multiple targets or objects. The association between measurements possible origins within range is difficult, especially for occlusion detection blind spots, which can lead missing measurements. This paper aims solve this challenge solving component in tracking. A multi algorithm based on Generalized maximum clique problem (GMMCP) proposed. transformed into a...