Rice has an important position in China as well the world. With wide application of rice hybridization technology, problem mixing between individual varieties become more and prominent, so variety identification is for agricultural production, phenotype collection, scientific breeding. Traditional methods are highly subjective time-consuming. To address this issue, we propose a novel locality preserved selective projection learning (LPSPL) method non-destructive based on leaf hyperspectral...

Information technology and statistical modeling have made significant contributions to smart agriculture. Machine vision hyperspectral technologies, with their non-destructive real-time capabilities, been extensively utilized in the diagnosis quality monitoring of crops seeds, becoming essential tools traditional This work applies these techniques address color classification rapeseed, which is great significance field rapeseed growth research. To bridge gap between machine technology, a...

Image denoising remains a fundamental challenge in digital image processing due to the inevitable presence of noise during acquisition and transmission. While existing filtering methods predominantly focus on local spatial information, they often overlook crucial structural information from other perspectives, such as manifold global structures. To address this limitation, we propose novel linear projection-based (LPNF) framework grounded projection learning theory. This innovatively learns...

Hyperspectral image (HSI) change detection (CD) is an important technology for identifying surface changes using multi-temporal HSIs. Nevertheless, the high dimensionality of HSIs presents significant challenges CD tasks, including issues such as lack robustness and computational costs in existing methods. To address those issues, this paper proposes unsupervised simple effective HSI model termed L2,1-norm regularized double non-negative matrix factorization (L2,1-DNMF). Specifically,...

Strawberry grading by picking robots can eliminate the manual classification, reducing labor costs and minimizing damage to fruit. size or weight is a key factor in grading, with accurate estimation being crucial for proper classification. In this paper, we collected 1521 sets of strawberry RGB-D images using depth camera manually measured strawberries construct training dataset regression model. To address issue incomplete caused environmental interference cameras, study proposes multimodal...

To improve the accuracy and reliability of orchard spraying robots, an integrated navigation system was developed, consisting a real-time kinematic positioning-Beidou satellite (RTK-BDS) receiver, inertial measurement unit (IMU), controller, servo motors. Using loose coupling combination method, error Kalman filter algorithm based on position heading angle is implemented to correct in real time. Combining kinematics model pure pursuit robot, path-tracking control proposed. Path planning...

Accurate prediction of soil properties is essential for sustainable land management and precision agriculture. This study presents an LSTM-CNN-Attention model that integrates temporal spatial feature extraction with attention mechanisms to improve predictive accuracy. Utilizing the LUCAS dataset, analyzes spectral data estimate key properties, including organic carbon (OC), nitrogen (N), calcium carbonate (CaCO3), pH (in H2O). The Long Short-Term Memory (LSTM) component captures...

Massive events can be produced today because of the rapid development Internet Things (IoT). Complex event processing, which used to extract high-level patterns from raw data, has become an essential part IoT middleware. Prediction analytics is important technology in supporting proactive complex processing. In this paper, we propose use dynamic Bayesian model averaging develop a high-accuracy prediction analytic method for large-scale application. This method, based on new multilayered...

Low-rank representation (LRR) is widely utilized in image feature extraction, as it can reveal the underlying correlation structure of data. However, subspace learning methods based on LRR suffer from problems lacking robustness and discriminability. To address these issues, this paper proposes a new robust extraction method named weighted Schatten p-norm minimization via low-rank discriminative embedding regression (WSNM-LRDER) method. This works by integrating linear into model. In...

Accurate and timely monitoring of pests is an effective way to minimize the negative effects in agriculture. Since deep learning-based methods have achieved good performance object detection, they been successfully applied for pest detection monitoring. However, current fail balance relationship between computational cost model accuracy. Therefore, this paper proposes a lightweight, locality-aware faster R-CNN (LLA-RCNN) method real-time The proposed uses MobileNetV3 replace original...

Hyperspectral images (HSIs) capture a wide range of spectral features across multiple bands light, from visible to near-infrared. image classification technology enables researchers accurately identify and analyze the composition distribution surface materials. Current mainstream deep learning methods typically use block sampling spatial for model. However, this approach can affect results due influence neighboring within sample block. To improve model’s focus on center block, study proposes...

Although linear discriminant analysis (LDA)-based subspace learning has been widely applied to hyperspectral image (HSI) classification, the existing LDA-based methods exhibit several limitations: (1) They are often sensitive noise and demonstrate weak robustness; (2) these ignore local information inherent in data; (3) number of extracted features is restricted by classes. To address drawbacks, this paper proposes a novel joint sparse (JSLLDA) method integrating embedding regression...

Watermelon is a crop susceptible to diseases. Rapid and effective detection of watermelon diseases great significance ensure the yield watermelon. Aiming at interference environment obstacles in natural environment, resulting low target accuracy poor robustness, this paper takes leaves as research object, considering anthracnose, leaf blight, spot normal examples. A disease recognition method based on deep learning proposed. This has improved pre-selected box setting formula SSD model tested...

The growth of real estate sector has been significantly influenced in recent years by ongoing regulation acquisition policy and the effects COVID-19 epidemic on economy. fluctuation housing prices is one most concerning factors for prospective homeowners. Whether property can sustain a comparatively constant level an extended period time crucial factor Numerous modeling application techniques prediction algorithms, together with promotion machine learning offer fresh approaches to forecast...

Distributed denial of service (DDoS) attack is one the major threats to current Internet. It challenging detect DDoS attacks accurately and quickly. We propose a novel IP Flow Interaction Feature algorithm (FIF) based on multiple features flows via addresses ports. To increase detection accuracy in various conditions, we describe state characteristics network using FIF time series, simple but efficient FIF-based model (FDAD) proposed by associating with contextual information observed...

<title>Abstract</title> To date, existing quiescent-state-based safe memory reclamation algorithms have relied on a global epoch counter maintained through software timestamps. While most approaches demonstrate good performance when employed in lock-free concurrent data structures, they grapple with issues concerning robustness. Specifically, interval-based (IBR), thread reclaiming shared block must compare whether the lifetime of intersects all threads’ reserved intervals. IBR can provide...

Image super-resolution has experienced significant advancements with the emergence of deep learning technology. However, deploying highly complex networks on resource-constrained devices poses a challenge due to their substantial computational requirements. This paper presents Adaptive Dynamic Shuffle Convolutional Parallel Network (ADSCPN), novel lightweight model designed achieve an optimal balance between efficiency and image reconstruction quality. The ADSCPN framework employs...

Multi-face alignment aims to identify geometry structures of multiple faces in an image, and its performance is essential for the many practical tasks, such as face recognition, tracking, animation. In this work, we present a fast bottom-up multi-face approach, which can simultaneously localize multi-person facial landmarks with high precision. more detail, our architecture maps high-dimensional space all are represented. By clustering features belonging same face, approach align...

We proposed a recovery scheme for image deblurring. The is under the framework of sparse representation and it has three main contributions. Firstly, considering property natural image, nonlocal overcompleted dictionaries are learned patches in our scheme. And, then, we coded each clustering with corresponding dictionary to recover whole latent image. In addition, some practical applications, also method evaluate blur kernel make algorithm usable blind recovery. experimental results...