Plant disease is one of the primary causes crop yield reduction. With development computer vision and deep learning technology, autonomous detection plant surface lesion images collected by optical sensors has become an important research direction for timely diagnosis. In this paper, anthracnose method based on proposed. Firstly, problem insufficient image data caused random occurrence apple diseases, in addition to traditional augmentation techniques, Cycle-Consistent Adversarial Network...

The failure of an insulator may compromise the safety entire power transmission system. Therefore, defect detection is vital for safe operation systems. However, defects in image have varying sizes, and several currently available methods do not satisfactory accuracy small defects. To address this issue, we propose improved network with a batch normalization convolutional block attention module (BN-CBAM) feature fusion module. BN-CBAM designed to better exploit channel information enhance...

In this paper, we design and build a power line inspection robot capable of hybrid operation modes. Specifically, the developed is able to land on overhead ground wire (OGW) move as climbing robot. When negotiate obstacles, it can vertically take off fly over obstacles unmanned aerial vehicle (UAV). A customized trumpet-shaped undercarriage used guarantee that safely. With aid swingable 2D Laser Range Finder (LRF), not only determine whether there are but also detect position orientation...

In recent decades, high-resolution (HR) remote sensing images have shown considerable potential for providing detailed information change detection. The traditional detection methods based on HR mostly only detect a single land type or the range, and cannot simultaneously of all object types pixel-level range changes in area. To overcome this difficulty, we propose new coarse-to-fine deep learning-based land-use method. We independently created scene classification dataset called NS-55,...

This paper presents a detection method of insulator stings for aerial inspection based on feature-fusion. The local sub-images strings are firstly collected from videos and tagged to establish training dataset. fusion feature is then composed by the histogram oriented gradients (HOG) binary pattern (LBP) after principal component analysis (PCA) dimension reduction separately. A model developed SVM algorithm with feature. At phase, threshold segmentation morphological operation adopted...

Abstract: Impaired osseointegration of the implant remains big hurdle for dental therapy in diabetic patients. In this study, authors first identified that miR204 was strikingly highly expressed bone mesenchymal stem cells (BMSCs) rats. Forced expression repressed osteogenic potential BMSCs, while inhibition significantly increased capacity. Moreover, inhibitor conjugated with gold nanoparticles (AuNP-antagomiR204) and dispersed them poly(lactic-co-glycolic acid) (PLGA) solution. The...

The power transmission mainly depends on overhead infrastructures, such as towers and lines. Automatic inspection by robots or UAVs for the infrastructures is an essential way to ensure safety of transmission. detection classification prerequisite automatic inspection. This paper compares two state-of-art deep learning methods realize high-voltage tower detection. We build dedicated dataset multi-object detection, including data collection, preprocessing annotation. After that, models...

Building Change Detection (BCD) is one of the core issues in earth observation and has received extensive attention recent years. With rapid development technology, data source remote sensing change detection continuously enriched, which provides possibility to describe spatial details ground objects more finely characterize with multiple perspectives levels. However, due different physical mechanisms multi-source data, BCD based on heterogeneous a challenge. Previous studies mostly focused...

Image-based segmentation of overhead power lines is critical for line inspection. Real-time helps the inspection robot avoid obstacles or land on wire during task. It challenging several studies to achieve real-time with high accuracy. In addition, cluttered background brings great difficulties segmentation. To address these issues, an efficient parallel branch network proposed. Our framework combines a context that generates useful global information spatial preserves high-resolution...

In this letter, we propose RI-LIO, a new reflectivity image assisted tightly-coupled LiDAR-inertial odometry (LIO) framework that introduces additional texture information to efficiently reduce the drift of geometric-only methods. To achieve this, construct an iterated extended Kalman filter by blending point-to-plane geometric measurement and measurement. Specifically, is defined as distance from raw point scan its nearest neighbor plane in global incremental kd-tree map. The searched used...

The conventional inspection of fragile insulators is critical to grid operation and insulator segmentation the basis inspection. However, various still difficult because great differences in colour shape, as well cluttered background. Traditional algorithms need many artificial thresholds, thereby limiting adaptability algorithms. A compact end‐to‐end neural network, which trained framework conditional generative adversarial networks, proposed for real‐time pixel‐level insulators. input...

There are a large number of insulators on the transmission line, and insulator damage will have major impact power supply security. Image-based segmentation in lines is premise also critical task for line inspection. In this paper, modified conditional generative adversarial network pixel-level proposed. The generator reconstructed by encoder-decoder layers with asymmetric convolution kernel which can simplify complexity extract more kinds feature information. discriminator composed fully...

In recent years, various types of inspection robots have been developed to automate powerline inspection. The hybrid robot combines the advantages climbing and flying has a promising prospect in But landing on target among multiple ones is challenging. Flights require robust detection powerlines stable tracking powerline. We propose complete solution for autonomous First, special feature extraction operator corresponding density-based recognition algorithm are designed detect multiscale...

This letter presents the first trajectory planning method for hybrid robot to perform powerline inspection involving obstacle navigation and landing. We develop a geometric model that incorporates constraints landing on powerline, avoidance, objectives maximize visibility of during flight. The generation is achieved via solving multiple shooting nonlinear programming problem with respect system dynamics constraints. formulation accommodates both powerline-to-powerline air-to-powerline...

Conventional face anti-spoofing methods might be poorly generalized to unseen data distributions. Thus, we improve the generalization of spoof detection from multi-domain feature disentanglement. Specially, a two-branch convolutional network is proposed separate spoof-specific features and domain-specific images explicitly. The are further used for live vs. classification. To minimize correlation among these two features, present cross-adversarial training scheme, which requires each branch...

The principle of multiresolution segmentation was represented in detail this study, and the canny algorithm applied for edge-detection a remotely sensed image based on principle. target divided into regions object-oriented edge-detection. Furthermore, object hierarchy created, series features (water bodies, vegetation, roads, residential areas, bare land other information) were extracted by spectral geometrical features. results indicate that has positive effect segmentation, overall...