Search Results (509)

Graphical abstract
Full article ">Figure 1
<p>A whole-cell biosensor integrated with bioluminescence.</p> Full article ">Figure 2
<p>Abscisic acid (ABA) interaction with other hormones (<b>A</b>) ABA response for abiotic stress response, stomatal closure, seed germination, root development, etc. (<b>B</b>) Interaction of ABA with MAPK.</p> Full article ">Figure 3
<p>Plant hormone interaction: (<b>a</b>) interaction of SnRK2, PP2C, and MAPK without ABA; and (<b>b</b>) interaction of SNRK2, PP2C, and MAPK in the presence of ABA.</p> Full article ">Figure 4
<p>The plot of ABA concentration (µmol) and bioluminescence (RLUs) shows proportionality between varying ABA concentration and resulting bioluminescence. The gradient from blue to orange on the plot represents bioluminescence intensity. Blue color is indicating the lower and orange color is indicating the higher levels, corresponding to ABA concentrations and production rates.</p> Full article ">

Figure 1
<p>Multipath routing in the megaconstellation network.</p> Full article ">Figure 2
<p>The overall structure of DRL-based multipath routing.</p> Full article ">Figure 3
<p>Track of sub-satellite point of megaconstellation networks.</p> Full article ">Figure 4
<p>The framework of the proposed GNN-based multipath traffic scheduling (GMTS) algorithm.</p> Full article ">Figure 5
<p>DRL-based multipath routing for LEO megaconstellation networks.</p> Full article ">Figure 6
<p>Average throughput on different constellations with various traffic intensity. (<b>a</b>) Average throughput for Iridium. (<b>b</b>) Average throughput for OneWeb.</p> Full article ">Figure 7
<p>Average flow completion ratio of different constellations with various traffic intensity. (<b>a</b>) Average flow completion ratio for Iridium constellation. (<b>b</b>) Average flow completion ratio for OneWeb constellation.</p> Full article ">Figure 8
<p>Average end-to-end delay for different constellations with various traffic intensities. (<b>a</b>) Average end-to-end delay for Iridium constellation. (<b>b</b>) Average end-to-end delay for OneWeb constellation.</p> Full article ">

Figure 1
<p>Distribution of objects with different orientations and process from density map to mask. (<b>a</b>,<b>b</b>) Objects with the same or similar orientations are distributed regionally; these regions are marked with yellow boxes. (<b>c</b>) The original RS image. (<b>d</b>) The object to be detected in the density map is highlighted, such as the large vehicle in (<b>c</b>), and the area with more objects has higher pixel intensity. (<b>e</b>) A mask is generated on the basis of the density map by setting a certain threshold, as shown with the yellow rectangles.</p> Full article ">Figure 2
<p>Architecture of DDE-Net. DDE-Net consists of three proposed modules (<b>a</b>): a backbone network (<b>b</b>), FPN (<b>c</b>), and FAM and ODM (<b>d</b>).</p> Full article ">Figure 3
<p>Density map generation network in DGM. It comprises three parallel CNNs, each with filters that have varying local receptive field sizes. Pooling is applied for each 2 × 2 region, and ReLU is adopted as the activation function.</p> Full article ">Figure 4
<p>Visualization of masks with different thresholds. A smaller threshold value will obtain a larger range of local object areas, while increasing the threshold value will increase the number of areas, but the area of each area will be reduced. The thresholds (<b>a</b>–<b>c</b>) are 0.001, 0.01, and 0.1.</p> Full article ">Figure 5
<p>Components in MRM. (<b>a</b>,<b>b</b>) Convolution layer and pooling layer to extract the feature. (<b>c</b>) Routing prediction in two branches.</p> Full article ">Figure 6
<p>AWC in SCM. Convolution kernels apply radial rotation transformation according to the rotation angle <math display="inline"><semantics> <mrow> <mo>[</mo> <msub> <mi>θ</mi> <mn>1</mn> </msub> <mo>,</mo> <mo>…</mo> <mo>,</mo> <msub> <mi>θ</mi> <mi>n</mi> </msub> <mo>]</mo> </mrow> </semantics></math>. The affine transformation matrix corresponding to the angle <math display="inline"><semantics> <mi>θ</mi> </semantics></math> rotates the original convolution kernel sampling points through the transformation matrix to obtain the new convolution kernel sampling points after rotation. The <span class="html-italic">n</span> rotated convolution kernels convolute the input x independently, and the features extracted by each convolution are weighted and summed according to the weight <math display="inline"><semantics> <mrow> <mo>[</mo> <msub> <mi>λ</mi> <mn>1</mn> </msub> <mo>,</mo> <mo>…</mo> <mo>,</mo> <msub> <mi>λ</mi> <mi>n</mi> </msub> <mo>]</mo> </mrow> </semantics></math>.</p> Full article ">Figure 7
<p>Ablation studies of different modules in DDE-Net. Vanilla DDE-Net refers to the network without DGM, MRM, and SCM, as illustrated in <a href="#electronics-13-03029-f002" class="html-fig">Figure 2</a>.</p> Full article ">Figure 8
<p>Visual comparison. The red box in the images indicates errors made by other methods.</p> Full article ">Figure 9
<p>Visualization results on DOTA.</p> Full article ">Figure 10
<p>Visualization results on HRSC2016.</p> Full article ">Figure 11
<p>Errors that may occur with DDE-Net.</p> Full article ">

Figure 1
<p>A UAV-assisted Internet of Vehicles (IoV) scenario.</p> Full article ">Figure 2
<p>UAV communication architecture.</p> Full article ">Figure 3
<p>UAV centralized communication.</p> Full article ">Figure 4
<p>UAV decentralized communication.</p> Full article ">Figure 5
<p>Categorization of resources for management in UAV-assisted IoV networks.</p> Full article ">Figure 6
<p>Joint resource management metrics in UAV-assisted IoV Networks.</p> Full article ">Figure 7
<p>Classification of routing protocols in UAV-IoV Networks.</p> Full article ">

Figure 1
<p>An example of the centroid of two depots.</p> Full article ">Figure 2
<p>An example of the exchange operator.</p> Full article ">Figure 3
<p>Average costs of each level and neighborhood structure.</p> Full article ">Figure 4
<p>Average execution times of each level and neighborhood structure.</p> Full article ">Figure 5
<p>Average total cost and distance for the entire algorithm execution.</p> Full article ">Figure 6
<p>Average execution time for the entire algorithm implementation.</p> Full article ">

Figure 1
<p>Illustration of the concepts of detection and classification.</p> Full article ">Figure 2
<p>Illustration of the architecture of a CNN.</p> Full article ">Figure 3
<p>Comparison of threshold methods.</p> Full article ">Figure 4
<p>Use of Adaptive Threshold to determine degradation status: (<b>a</b>) Outcome for a crosswalk exhibiting signs of wear; (<b>b</b>) Outcome for a crosswalk without signs of wear; (<b>c</b>) Outcome for a crosswalk exhibiting a considerable degree of wear.</p> Full article ">Figure 5
<p>Example of a question asked on the form.</p> Full article ">Figure 6
<p>Example of the results obtained in the questionnaire.</p> Full article ">Figure 7
<p>Example of how the increase in pixels can influence the results.</p> Full article ">Figure 8
<p>Illustration from CNN CSPDarknet53-tiny. Adapted from [<a href="#B9-applsci-14-06462" class="html-bibr">9</a>].</p> Full article ">Figure 9
<p>Architecture of the YOLOv4-tiny model. Adapted from [<a href="#B11-applsci-14-06462" class="html-bibr">11</a>].</p> Full article ">Figure 10
<p>Example of a captured image.</p> Full article ">Figure 11
<p>Illustration of objects captured by a 90° camera vs. a 160° camera.</p> Full article ">Figure 12
<p>Diagram of how the first approach works.</p> Full article ">Figure 13
<p>Examples of Adaptive Threshold detection: (<b>a</b>) Bounding boxes of three crosswalks (1)–(3) detected by the YOLOv4-tiny model; (<b>b</b>) Result of using the Adaptive Threshold method to determine the degradation state of the three crosswalks (1)–(3).</p> Full article ">Figure 14
<p>Examples of Adaptive Threshold detection in scenes without shadows: (<b>a</b>) Example 1: detection (1), identification (2), and classification (3) process; (<b>b</b>) Example 2: detection (1), identification (2), and classification (3) process.</p> Full article ">Figure 15
<p>Comparison between the two approaches in two different examples: Example (<b>a</b>): (1) and (2) shows the detection using the first approach considered—YOLOv4-tiny and Adaptive Threshold. (3) shows the result obtained under the same conditions but using the second approach—YOLOv4-tiny with three classes. Example (<b>b</b>): (1) and (2) shows the detection using the first approach considered—YOLOv4-tiny and Adaptive Threshold. (3) shows the result obtained under the same conditions but using the second approach—YOLOv4-tiny with three classes.</p> Full article ">Figure 16
<p>The architecture of the prototype was developed.</p> Full article ">Figure 17
<p>Schematic of the electrical circuit of the various hardware components.</p> Full article ">Figure 18
<p>Hardware components of the prototype developed.</p> Full article ">Figure 19
<p>Use case diagram.</p> Full article ">Figure 20
<p>Flowchart of how the Raspberry Pi works.</p> Full article ">Figure 21
<p>Structure required in documents.</p> Full article ">Figure 22
<p>Endpoints on the API.</p> Full article ">Figure 23
<p>Example of a response from the Geoapify API.</p> Full article ">Figure 24
<p>Sequence diagram between API and Raspberry PI.</p> Full article ">Figure 25
<p>Illustration of the mechanism for identifying the distance between coordinates.</p> Full article ">Figure 26
<p>Sequence diagram of communication between the API and the CrosSafe application.</p> Full article ">Figure 27
<p>Mockup Home page.</p> Full article ">Figure 28
<p>Mockup Dashboard page.</p> Full article ">Figure 29
<p>Crosswalk counting component by state of degradation.</p> Full article ">Figure 30
<p>Interactive map in the application.</p> Full article ">Figure 31
<p>Table with additional information on crosswalks detected using the pagination method.</p> Full article ">Figure 32
<p>Popup for confirmation of crosswalk repair.</p> Full article ">Figure 33
<p>Example of a success message following confirmation of a crosswalk repair.</p> Full article ">Figure 34
<p>An example of a crosswalk classified as severely worn.</p> Full article ">Figure 35
<p>Example of the number of documents received by Firestore.</p> Full article ">Figure 36
<p>Illustration of the route taken.</p> Full article ">Figure 37
<p>Detection and classification time required.</p> Full article ">Figure 38
<p>An image containing a crosswalk slightly diagonally.</p> Full article ">Figure 39
<p>Visualization of the data transmitted by the Raspberry PI in the CrosSafe web application.</p> Full article ">

Figure 1
<p>Hybrid DCTE strategy processing flowchart.</p> Full article ">Figure 2
<p>Three robots obtained optimal path after decentralized path planning step.</p> Full article ">Figure 3
<p>Red robot and blue robot re-planned the trajectory to avoid the collision.</p> Full article ">Figure 4
<p>After replanning the path, the optimal paths are shown in the green block.</p> Full article ">Figure 5
<p>Path re-optimization and connection.</p> Full article ">Figure 6
<p>Three robots plan a path after decentralized path planning step, based on DQN.</p> Full article ">Figure 7
<p>Robot 1 and robot 3 replan the optimal paths simultaneously, based on the DQN RL algorithm.</p> Full article ">Figure 8
<p>After path re-optimization and connection step, the optimized paths for robot 1, 2, and 3, respectively.</p> Full article ">Figure 9
<p>Optimal paths for robots 1, 2, and 3, based on DQN-based approach from ref [<a href="#B22-electronics-13-02927" class="html-bibr">22</a>].</p> Full article ">Figure 10
<p>(<b>a</b>) Gazebo simulator environment; (<b>b</b>) Agilex LIMO robot.</p> Full article ">Figure 11
<p>Three robots explore and track the optimal paths, based on a hybrid DCTE strategy: (<b>a</b>) t = 0 s; (<b>b</b>) t = 39 s; (<b>c</b>) t = 66 s; (<b>d</b>) t = 96 s; (<b>e</b>) t = 133 s; (<b>f</b>) t = 162 s; (<b>g</b>) t = 201 s; (<b>h</b>) t = 252 s; (<b>i</b>) t = 308 s.</p> Full article ">Figure 12
<p>Three reach their target points efficiently, quickly, and safely, in a complex environment.</p> Full article ">

Figure 1
<p>Six different kinds of PCB defects.</p> Full article ">Figure 2
<p>The overall network architecture of the proposed local and global context-enhanced lightweight CenterNet.</p> Full article ">Figure 3
<p>Ratio of PCB defect bounding box area to total image area.</p> Full article ">Figure 4
<p>The structure of local coordinate attention and global self-attention.</p> Full article ">Figure 5
<p>Sparse attention is used to skip computations in the most irrelevant region, and pooling is used to downsample the key and value to reduce FLOPs.</p> Full article ">Figure 6
<p>Detection results of different object detection algorithms. More detection results of the other defects can be found in the <a href="#app1-sensors-24-04729" class="html-app">Supplementary Materials</a>.</p> Full article ">

Figure 1
<p>SD-GPSR routing architecture.</p> Full article ">Figure 2
<p>The process of SD-GPSR handling packets.</p> Full article ">Figure 3
<p>Calculation of relay node angle.</p> Full article ">Figure 4
<p>Division of the communication range of forwarding nodes.</p> Full article ">Figure 5
<p>Comparison of the influence of four protocols on network performance with data service changes.</p> Full article ">Figure 6
<p>Comparison of control costs in different scenarios.</p> Full article ">Figure 7
<p>Comparison of average packet loss rate in different scenarios.</p> Full article ">Figure 8
<p>Comparison of average end-to-end delay in different scenarios.</p> Full article ">

Figure 1
<p>Detailed architecture of the proposed DCP-YOLOv8.</p> Full article ">Figure 2
<p>Structure diagram of the Deformable Convolutions v3.</p> Full article ">Figure 3
<p>Structure of the C2f_DCNv3.</p> Full article ">Figure 4
<p>Structure of the RCSP.</p> Full article ">Figure 5
<p>Structure of the DCNv3-Dyhead.</p> Full article ">Figure 6
<p>Distribution of the number of defective samples in the training set.</p> Full article ">Figure 7
<p>Analysis during training; (<b>a</b>) represents the box loss curve, (<b>b</b>) represents the classification loss curve, and (<b>c</b>) represents the map curve changes of YOLOv8n and DCP-YOLOv8.</p> Full article ">Figure 8
<p>Some examples of defect detection effects.</p> Full article ">Figure 8 Cont.
<p>Some examples of defect detection effects.</p> Full article ">Figure 9
<p>Robustness analysis of DCP-YOLOv8 in complex environments.</p> Full article ">Figure 9 Cont.
<p>Robustness analysis of DCP-YOLOv8 in complex environments.</p> Full article ">Figure 10
<p>Some examples of defective target heat maps.</p> Full article ">Figure 10 Cont.
<p>Some examples of defective target heat maps.</p> Full article ">

Figure 1
<p>Illustration of and comparison between different stream-ordering methods. (<b>a</b>) Lower limits of the feasible stream-order values from our proposed methods. (<b>b</b>) The upper limits of the feasible stream-order values from our proposed methods. (<b>c</b>) Another feasible set of stream-order values conforming to the two rules of our proposed ordering methods, as described in <a href="#water-16-01965-t001" class="html-table">Table 1</a>. (<b>d</b>) The Strahler stream-ordering method. (<b>e</b>) The Shreve stream-ordering method.</p> Full article ">Figure 2
<p>Yangtze River basin network delineated from the MERIT-Hydro dataset. The black lines outline the national boundaries or coastal lines of China, providing a geographical context. The color-coded lines depict the river reaches. Each color indicates the total number of reaches, encompassing the current reach and all upstream reaches.</p> Full article ">Figure 3
<p>Diagram of workflow.</p> Full article ">Figure 4
<p>Spatial distribution of stream orders over the Yangtze River basin. (<b>a</b>) The Strahler order. (<b>b</b>) The Shreve order. (<b>c</b>) The lower bounds of the proposed order family. (<b>d</b>) The upper bounds of the proposed order family.</p> Full article ">Figure 5
<p>The count of stream-order values with a limited number of reaches. For the sake of clarity, only stream orders with 16 or fewer reaches are illustrated.</p> Full article ">Figure 6
<p>Parallelization speedup (<b>a</b>) and efficiency (<b>b</b>) for the Shreve orders and the proposed stream orders, respectively. The solid black lines represent the ideal speedup or efficiency under conditions of perfect scaling. The dashed black lines indicate the limits imposed by the upstream-to-downstream sequentiality along the longest flow path.</p> Full article ">Figure 7
<p>Computation deficiencies observed for routing each individual reach at a processor count of 51 for the Yangtze River network.</p> Full article ">Figure 8
<p>Histogram of computation deficiencies observed for the Yangtze River network at a processor count of 51.</p> Full article ">Figure 9
<p>Optimization of the stream order. (<b>a</b>) Decrease in the total computation time with the optimization iteration. (<b>b</b>) Comparison of the histogram of computation deficiencies between the optimized stream orders and the upper bounds of the stream-order family. (<b>c</b>) Spatial distribution of the computation deficiencies.</p> Full article ">Figure 10
<p>Same as <a href="#water-16-01965-f006" class="html-fig">Figure 6</a> but for the optimized stream orders and the upper bounds of the proposed order family. (<b>a</b>) Parallelization speedup. (<b>b</b>) Parallelization efficiency.</p> Full article ">Figure 11
<p>Same as <a href="#water-16-01965-f004" class="html-fig">Figure 4</a>d but for the stream orders with one (<b>a</b>) and three (<b>b</b>) breaks along the longest flow path. The red dots denote the breakpoints. The halo reaches are not shown for clarity.</p> Full article ">Figure 12
<p>Same as <a href="#water-16-01965-f006" class="html-fig">Figure 6</a> but for the upper bounds of the order family with two (the orange lines) or four (the blue lines) breaks along the longest flow path. (<b>a</b>) Parallelization speedup. (<b>b</b>) Parallelization efficiency.</p> Full article ">

Figure 1
<p>A typical route of a vehicle.</p> Full article ">Figure 2
<p>Illustration of time windows and the latest response time.</p> Full article ">Figure 3
<p>Interaction diagram of the agents.</p> Full article ">Figure 4
<p>Flowchart of the agent algorithms.</p> Full article ">Figure 5
<p>A numerical example of the proposed interchange mechanism (<b>a</b>) search for a feasible vehicle, (<b>b</b>) search for a vehicle with profitable assignment.</p> Full article ">Figure 6
<p>The impact of congestion factor on performance measures (<b>a</b>) unit cost and unit profit under interchange mechanism levels, (<b>b</b>) order acceptance ratio and loaded trip ratio under interchange mechanism levels, (<b>c</b>) unit cost and unit profit under multiple trip order ratio levels, (<b>d</b>) order acceptance ratio and loaded trip ratio under multiple trip order ratio levels.</p> Full article ">Figure 6 Cont.
<p>The impact of congestion factor on performance measures (<b>a</b>) unit cost and unit profit under interchange mechanism levels, (<b>b</b>) order acceptance ratio and loaded trip ratio under interchange mechanism levels, (<b>c</b>) unit cost and unit profit under multiple trip order ratio levels, (<b>d</b>) order acceptance ratio and loaded trip ratio under multiple trip order ratio levels.</p> Full article ">Figure 7
<p>The impact of response time on performance measures (<b>a</b>) unit cost and unit profit under congestion levels, (<b>b</b>) order acceptance ratio and loaded trip ratio under congestion levels, (<b>c</b>) unit cost and unit profit under interchange mechanism levels, (<b>d</b>) order acceptance ratio and loaded trip ratio under interchange mechanism levels, (<b>e</b>) unit cost and unit profit under multiple trip order ratio levels, (<b>f</b>) order acceptance ratio and loaded trip ratio under multiple trip order ratio levels.</p> Full article ">Figure 7 Cont.
<p>The impact of response time on performance measures (<b>a</b>) unit cost and unit profit under congestion levels, (<b>b</b>) order acceptance ratio and loaded trip ratio under congestion levels, (<b>c</b>) unit cost and unit profit under interchange mechanism levels, (<b>d</b>) order acceptance ratio and loaded trip ratio under interchange mechanism levels, (<b>e</b>) unit cost and unit profit under multiple trip order ratio levels, (<b>f</b>) order acceptance ratio and loaded trip ratio under multiple trip order ratio levels.</p> Full article ">Figure 8
<p>The impact of multiple-trip order ratio on performance measures (<b>a</b>) unit cost and unit profit under interchange mechanism levels, (<b>b</b>) order acceptance ratio and loaded trip ratio under interchange mechanism levels.</p> Full article ">Figure 9
<p>The impact of the problem size on the CPU time.</p> Full article ">

Figure 1
<p>Proposed secure communication scheme.</p> Full article ">Figure 2
<p>Key exchange and authentication process.</p> Full article ">Figure 3
<p>MITM attack detection.</p> Full article ">Figure 4
<p>Results through OFMC and AtSe.</p> Full article ">Figure 5
<p>Experimental setup of NFV-based IPv4-IPv6 virtual networks.</p> Full article ">Figure 6
<p>Server output.</p> Full article ">Figure 7
<p>Client output.</p> Full article ">Figure 8
<p>Connectivity and traffic path.</p> Full article ">

Figure 1
<p>PRISMA Flowchart. Self-elaboration based on Scopus.</p> Full article ">Figure 2
<p>Articles published per year on the topic of ISDN through the use of AI. Source. Self-elaboration based on Scopus and Bibliometrix.</p> Full article ">Figure 3
<p>Articles published by authors on the topic of ISDN through the use of AI. Source. Self-elaboration based on Scopus and Bibliometrix.</p> Full article ">Figure 4
<p>Trend and evolution of keywords over time in the theme of ISDN through the use of AI. Source. Self-elaboration based on Scopus and Bibliometrix.</p> Full article ">Figure 5
<p>Correlation of keywords in the theme of ISDN through the use of AI. Source. Self-elaboration based on Scopus and VosViewer.</p> Full article ">Figure 6
<p>Relevance of keywords in the theme of ISDN through the use of AI. Source. Self-elaboration based on Scopus and Bibliometrix.</p> Full article ">Figure 7
<p>SDN Architecture. Source: Self-elaboration based on literature review.</p> Full article ">Figure 8
<p>Relevance of Keywords in the ISDN Theme Based on the Use of AI. Source: Own elaboration based on Scopus and Bibliometrix.</p> Full article ">

Figure 1
<p>The signage utilized along the trails.</p> Full article ">Figure 2
<p>Examples of marked places.</p> Full article ">Figure 3
<p>An illustration of the scenario.</p> Full article ">Figure 4
<p>Prototype architecture.</p> Full article ">Figure 5
<p>The IoT device from the prototype’s sensor node.</p> Full article ">Figure 6
<p>Convolutional neural network structure.</p> Full article ">Figure 7
<p>Classification of one-stage detection and two-stage detection model types in object detection.</p> Full article ">Figure 8
<p>One-stage detection model architecture.</p> Full article ">Figure 9
<p>Two-stage detection model architecture.</p> Full article ">Figure 10
<p>3 × 3 grid division.</p> Full article ">Figure 11
<p>Architecture of YOLOv3-Tiny model.</p> Full article ">Figure 12
<p>Representation of bounding boxes for each class.</p> Full article ">Figure 13
<p>(<b>a</b>) Training loss and mAP of first training using YOLOv3-Tiny. (<b>b</b>) Training loss and mAP of new training using YOLOv3-Tiny.</p> Full article ">Figure 14
<p>A flowchart of the detection and classification algorithm.</p> Full article ">Figure 15
<p>Sensor node local database.</p> Full article ">Figure 16
<p>Hybrid cryptosystem.</p> Full article ">Figure 17
<p>Connection to Bluetooth service and exchange of RSA public keys.</p> Full article ">Figure 18
<p>Process of preparing to transmit data.</p> Full article ">Figure 19
<p>An illustration of a scenario in which a user using a bicycle passes within range of the Bluetooth signal from the sensor node.</p> Full article ">Figure 20
<p>Example of block of 10 records in JSON.</p> Full article ">Figure 21
<p>Communication between nodes when transmitting data.</p> Full article ">Figure 22
<p>Mobile application screen.</p> Full article ">Figure 23
<p>The output on the mobile application with information on the route started and the location of the nodes.</p> Full article ">Figure 24
<p>Output of data exchange between nodes.</p> Full article ">Figure 25
<p>Bridge node local database.</p> Full article ">Figure 26
<p>Output from synchronization service.</p> Full article ">Figure 27
<p>A flowchart describing how the bridge node application works.</p> Full article ">Figure 28
<p>Central database.</p> Full article ">Figure 29
<p>GET method for obtaining information from sensor nodes.</p> Full article ">Figure 30
<p>POST method for sending and creating records.</p> Full article ">Figure 31
<p>Login page of web application.</p> Full article ">Figure 32
<p>Main page of web application.</p> Full article ">Figure 33
<p>A web application page with a view of the trails and sensor nodes.</p> Full article ">Figure 34
<p>A web application page with information on the sensor node selected by the user.</p> Full article ">Figure 35
<p>Filtering by time period in a sensor node.</p> Full article ">Figure 36
<p>Example of detection with YOLOv3-Tiny model trained with improved dataset.</p> Full article ">Figure 37
<p>Second example of detection with YOLOv3-Tiny model trained with improved dataset.</p> Full article ">Figure 38
<p>Positioning the sensor node on the ground.</p> Full article ">Figure 39
<p>Part of the trail captured by the sensor node.</p> Full article ">Figure 40
<p>SSH connection to the sensor node via the Terminus application.</p> Full article ">Figure 41
<p>Classification and detection of motorcycle class.</p> Full article ">Figure 42
<p>Classification and detection of bicycle class.</p> Full article ">Figure 43
<p>Simultaneous identification of person and bicycle classes.</p> Full article ">Figure 44
<p>Simultaneous identification of person class.</p> Full article ">Figure 45
<p>The records in the sensor node database.</p> Full article ">Figure 46
<p>Starting the trail and connection establishment.</p> Full article ">Figure 47
<p>Data exchange between nodes and initialization of synchronization service.</p> Full article ">Figure 48
<p>Activating Wi-Fi on the Android device.</p> Full article ">Figure 49
<p>Data synchronized with central database.</p> Full article ">Figure 50
<p>Records in central database.</p> Full article ">Figure 51
<p>Class counts in PRN-3, filtered by detection dates.</p> Full article ">Figure 52
<p>The class counts in PRN-3 on the page showing the map and the sensor nodes distributed on the ground.</p> Full article ">Figure 53
<p>The class counts in PRN-3 filtered by detection dates on the page showing the map and the sensor nodes distributed on the ground.</p> Full article ">