Submit to Forests Review for Forests Propose a Special Issue

Journal Menu

Journal Browser

Electronics, Close-Range Sensors and Artificial Intelligence in Forestry

Print Special Issue Flyer
Special Issue Editors
Special Issue Information
Keywords
Published Papers

A special issue of Forests (ISSN 1999-4907). This special issue belongs to the section "Forest Inventory, Modeling and Remote Sensing".

Deadline for manuscript submissions: closed (12 June 2022) | Viewed by 48356

Printed Edition Available!
A printed edition of this Special Issue is available here.

Special issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry book cover image

Share This Special Issue

Special Issue Editors

Prof. Dr. Stelian Alexandru Borz

E-Mail Website
Guest Editor

Department of Forest Engineering, Forest Management Planning and Terrestrial Measurements, Faculty of Silviculture and Forest Engineering, Transilvania University of Brasov, Şirul Beethoven 1, 500123 Brasov, Romania
Interests: wood and biomass supply chain optimization; sensor technology; transport optimization; forest planning
Special Issues, Collections and Topics in MDPI journals

Dr. Andrea R. Proto

E-Mail Website
Guest Editor

Department of Agraria, Mediterranean University of Reggio Calabria, 89122 Reggio Calabria, Italy
Interests: forest mechanization; productivity; NDT evaluation and wood quality; measuring wood properties; wood technology; wood engineering; urban forestry; agro-forestry biomass; sustainable agro-forestry management
Special Issues, Collections and Topics in MDPI journals

Dr. Robert Keefe

E-Mail Website
Guest Editor

College of Natural Resources, 875 Perimeter Drive MS 1133, University of Idaho Experimental Forest, Moscow, ID 83844-1133, USA

Prof. Dr. Mihai Nita

E-Mail Website
Guest Editor

Department of Forest Engineering, Universitatea Transilvania Brasov, Braşov, Romania
Interests: remote sensing; GIS; forest and water; forest management; machine learning
Special Issues, Collections and Topics in MDPI journals

Special Issue Information

Dear colleagues,

The use of electronics, close-range sensing and artificial intelligence has changed the management paradigm in many of the current industries, in which big-data analytics by automated processes has become the backbone of decision making and improvement. Acknowledging the integration of electronics, devices, sensors and intelligent algorithms in many of the equipment used in forest operations, as well as their use in various forestry-related applications, reality is showing us that many of the applications of forest engineering still rely on data collected traditionally, which is resource-intensive, demanding analysis and, in many cases, being limited to establishing the specific behaviors of forest product systems and wood supply chain. This situation is often preventing from developing solutions for improvement or to accurately infer the laws behind the operation and management of such systems. In particular, partly mechanized systems, environmental impact assessment and ergonomics of operation are components or disciplines in which improvement is crucial and which could rely on new advancements in algorithms, computers, electronics and sensor science to find solutions, by robust solutions, to these pressing needs. This Special Issue is intended to cover the whole range of applications, as is typical to forestry and forest engineering in which one or more of the following could be addressed by high-quality research or review papers:

Moving from traditional to computer-aided time consumption and productivity studies and models, and from traditional to advanced methods to estimate forest growth, production, dynamics and disturbance;
Enhancing big-data collection, analysis and augmentation using integrated or external electronics, computers and sensor systems;
Development, validation, and use of human and forestry equipment activity recognition models developed using smart watches, smart phones, micro-electric sensors and Internet-of-Things devices;
Applications of virtual reality, teleoperation, telematics, and robotics in forestry;
Application of close-range sensing to solve important problems in forestry and forest operations, such as those related to productivity, ergonomics and environmental impact assessment;
Development and/or implementation of intelligent algorithms and artificial intelligence (AI) for multivariate analyses, classification and event-based management in forestry and forest engineering;
Adapting and implementing low-cost solutions to answer the pressing problems in the wood supply chain and to overcome bottlenecks;
Intelligent methods, devices, equipment, machines, protocols, processes and robots that have a direct application or use in forestry;
Effects of automation on the environment and human welfare.

Prof. Dr. Stelian Alexandru Borz
Dr. Andrea Rosario Proto
Dr. Robert Keefe
Dr. Mihai Nita
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Forests is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

forest operations
productivity
ergonomics
environmental impact assessment
close-range
sensors
big data
virtual reality
teleoperation
automation
artificial intelligence
forestry 4.0

Published Papers (13 papers)

Download All Papers

Order results

Result details

Show export options Show export options

Select all

Export citation of selected articles as:

Editorial

Jump to: Research

3 pages, 663 KiB

Open AccessEditorial

Electronics, Close-Range Sensors and Artificial Intelligence in Forestry

by Stelian Alexandru Borz, Andrea Rosario Proto, Robert Keefe and Mihai Daniel Niţă

Forests 2022, 13(10), 1669; https://doi.org/10.3390/f13101669 - 11 Oct 2022

Cited by 1 | Viewed by 1064

Abstract

The use of electronics, close-range sensing and artificial intelligence has changed the management paradigm in many of the current industries in which big data analytics by automated processes has become the backbone of decision making and improvement [...] Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

Research

Jump to: Editorial

23 pages, 10331 KiB

Open AccessArticle

Research on Tree Ring Micro-Destructive Detection Technology Based on Digital Micro-Drilling Resistance Method

by Xueyang Hu, Yili Zheng, Da Xing and Qingfeng Sun

Forests 2022, 13(7), 1139; https://doi.org/10.3390/f13071139 - 19 Jul 2022

Cited by 3 | Viewed by 1621

Abstract

Micro-drilling resistance method is a widely used tree ring micro-destructive detection technology. To solve the problem that the detection signal of the analog micro-drilling resistance method has excessive noise interference and cannot intuitively identify tree ring information, this research proposes a digital micro-drilling resistance method and provides a recommended hardware implementation. The digital micro-drilling resistance method adopts the photoelectric encoder instead of ADC as the signal sampling module. Through the theoretical analysis of the DC motor characteristic, the PWM closed-loop speed control, the detection principle of the digital method is given. Additionally, the experimental equipment that can complete the detection of the digital method and the analog method simultaneously is designed to carry out comparative experiments. The experimental results show that: (1) The detection results of the digital method have a better-quality signal which can intuitively identify the tree rings. (2) The average correlation coefficient reaches 0.9365 between the detection results of the digital method and the analog method. (3) The average Signal-to-Noise Ratio (SNR) of the digital method is 39.0145 dB, which is 19.2590 dB higher than that of the analog method. The average noise interference energy in the detection result of the digital method is only 1.27% of the analog method. In summary, hardware implementation of the digital micro-drilling resistance method can correctly reflect the tree ring information and significantly improve the signal quality of the micro-drilling resistance technology. This research is helping to improve the identification accuracy of micro-drilling resistance technology, and to develop the application of tree ring micro-destructive detection technology in the high-precision field. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

24 pages, 10745 KiB

Open AccessArticle

Development of a Robust Machine Learning Model to Monitor the Operational Performance of Fixed-Post Multi-Blade Vertical Sawing Machines

by Stelian Alexandru Borz, Gabriel Osei Forkuo, Octavian Oprea-Sorescu and Andrea Rosario Proto

Forests 2022, 13(7), 1115; https://doi.org/10.3390/f13071115 - 15 Jul 2022

Cited by 3 | Viewed by 1479

Abstract

Monitoring the operational performance of the sawmilling industry has become important for many applications including strategic and tactical planning. Small-scale sawmilling facilities do not hold automatic production management capabilities mainly due to using obsolete technology which is an effect of low financial capacity and focus their strategy on increasing value recovery and saving resources and energy. Based on triaxial acceleration data collected over five days at a sampling rate of 1 Hz, a robust machine learning model was developed with the purpose of using it to infer the operational events based on lower sampling rates adopted as a strategy to collect long-term data. Among its performance metrics, the model was characterized in its training phase by a very high overall classification accuracy (CA = 98.7%), F1 score (98.4%) and a very low error rate (LOG LOSS = 5.6%). For a three-class problem, it worked very well in classifying the main events related to the operation of the machine, with active work being characterized by an F1 score of 99.6% and an error of 3.6%. By accounting for the same metrics, the model was proven to be invariant to the sampling rates of up to 0.05 Hz (20 s) and produced even better results in the testing phase (CA = 98.9%, F1 = 98.6%, LOG LOSS = 5.5%, for a testing sample extracted at 0.05 Hz), while there were no differences in the share of class data irrespective of the sampling rate. The developed model not only preserves a high classification performance in the training and testing phases but it also seems to be invariant to lower sampling rates, making it useful for prediction over data collected at low sampling rates. In turn, this would enable the use of cheap data collectors to be operated for extended periods of time in various locations and will save human resources and money associated with data collection. Further tests would be required only for validation and they could be supported by collecting and feeding new data to the model to infer the long-term performance of similar sawmilling machines. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

19 pages, 4942 KiB

Open AccessArticle

Potential of Measure App in Estimating Log Biometrics: A Comparison with Conventional Log Measurement

by Stelian Alexandru Borz, Jenny Magaly Morocho Toaza, Gabriel Osei Forkuo and Marina Viorela Marcu

Forests 2022, 13(7), 1028; https://doi.org/10.3390/f13071028 - 30 Jun 2022

Cited by 8 | Viewed by 1769

Abstract

Wood measurement is an important process in the wood supply chain, which requires advanced solutions to cope with the current challenges. Several general-utility measurement options have become available by the developments in LiDAR or similar-capability sensors and Augmented Reality. This study tests the accuracy of the Measure App developed by Apple, running by integration into Augmented Reality and LiDAR technologies, in estimating the main biometrics of the logs. In a first experiment (E1), an iPhone 12 Pro Max running the Measure App was used to measure the diameter at one end and the length of 267 spruce logs by a free-eye measurement approach, then reference data was obtained by taking conventional measurements on the same logs. In a second experiment (E2), an iPhone 13 Pro Max equipped with the same features was used to measure the diameter at one end and the length of 200 spruce logs by a marking-guided approach, and the reference data was obtained similar to E1. The data were compared by a Bland and Altman analysis which was complemented by the estimation of the mean absolute error (MAE), root mean squared error (RMSE) and normalized root mean square error (NRMSE). In E1, nearly 86% of phone-based log diameter measurements were within ±1 cm compared to the reference data, of which 37% represented a perfect match. Of the phone-based log length measurements, 94% were within ±5 cm compared to the reference data, of which approximately 22% represented a perfect match. MAE, RMSE, and NRMSE of the log diameter and length were of 0.68, 0.96, and 0.02 cm, and of 1.81, 2.55, and 0.10 cm, respectively. Results from E2 were better, with 95% of the phone-based log diameter agreeing within ±1 cm, of which 44% represented a perfect match. As well, 99% of the phone-based length measurements were within ±5 cm, of which approximately 27% were a perfect match. MAE, RMSE, and NRMSE of the log diameter and length were of 0.65, 0.92, and 0.03 cm, and 1.46, 1.93, and 0.04 cm, respectively. The results indicated a high potential of replacing the conventional measurements for non-piled logs of ca. 3 m in length, but the applicability of phone-based measurement could be readily extended to log-end diameter measurement of the piled wood. Further studies could check if the accuracy of measurements would be enhanced by larger samples and if the approach has good replicability. Finding a balance between capability and measurement accuracy by extending the study to longer log lengths, different species and operating conditions would be important to characterize the technical limitations of the tested method. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

11 pages, 2862 KiB

Open AccessFeature PaperArticle

Toward a Unified TreeTalker Data Curation Process

by Enrico Tomelleri, Luca Belelli Marchesini, Alexey Yaroslavtsev, Shahla Asgharinia and Riccardo Valentini

Forests 2022, 13(6), 855; https://doi.org/10.3390/f13060855 - 30 May 2022

Cited by 5 | Viewed by 2677

Abstract

The Internet of Things (IoT) development is revolutionizing environmental monitoring and research in macroecology. This technology allows for the deployment of sizeable diffuse sensing networks capable of continuous monitoring. Because of this property, the data collected from IoT networks can provide a testbed for scientific hypotheses across large spatial and temporal scales. Nevertheless, data curation is a necessary step to make large and heterogeneous datasets exploitable for synthesis analyses. This process includes data retrieval, quality assurance, standardized formatting, storage, and documentation. TreeTalkers are an excellent example of IoT applied to ecology. These are smart devices for synchronously measuring trees’ physiological and environmental parameters. A set of devices can be organized in a mesh and permit data collection from a single tree to plot or transect scale. The deployment of such devices over large-scale networks needs a standardized approach for data curation. For this reason, we developed a unified processing workflow according to the user manual. In this paper, we first introduce the concept of a unified TreeTalker data curation process. The idea was formalized into an R-package, and it is freely available as open software. Secondly, we present the different functions available in “ttalkR”, and, lastly, we illustrate the application with a demonstration dataset. With such a unified processing approach, we propose a necessary data curation step to establish a new environmental cyberinfrastructure and allow for synthesis activities across environmental monitoring networks. Our data curation concept is the first step for supporting the TreeTalker data life cycle by improving accessibility and thus creating unprecedented opportunities for TreeTalker-based macroecological analyses. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

21 pages, 11047 KiB

Open AccessArticle

Design and Testing of a Novel Unoccupied Aircraft System for the Collection of Forest Canopy Samples

by Sean Krisanski, Mohammad Sadegh Taskhiri, James Montgomery and Paul Turner

Forests 2022, 13(2), 153; https://doi.org/10.3390/f13020153 - 20 Jan 2022

Cited by 8 | Viewed by 2920

Abstract

Unoccupied Aircraft Systems (UAS) are beginning to replace conventional forest plot mensuration through their use as low-cost and powerful remote sensing tools for monitoring growth, estimating biomass, evaluating carbon stocks and detecting weeds; however, physical samples remain mostly collected through time-consuming, expensive and potentially dangerous conventional techniques. Such conventional techniques include the use of arborists to climb the trees to retrieve samples, shooting branches with firearms from the ground, canopy cranes or the use of pole-mounted saws to access lower branches. UAS hold much potential to improve the safety, efficiency, and reduce the cost of acquiring canopy samples. In this work, we describe and demonstrate four iterations of 3D printed canopy sampling UAS. This work includes detailed explanations of designs and how each iteration informed the design decisions in the subsequent iteration. The fourth iteration of the aircraft was tested for the collection of 30 canopy samples from three tree species: eucalyptus pulchella, eucalyptus globulus and acacia dealbata trees. The collection times ranged from 1 min and 23 s, up to 3 min and 41 s for more distant and challenging to capture samples. A vision for the next iteration of this design is also provided. Future work may explore the integration of advanced remote sensing techniques with UAS-based canopy sampling to progress towards a fully-automated and holistic forest information capture system. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

18 pages, 2270 KiB

Open AccessFeature PaperArticle

Estimating Aboveground Biomass in Dense Hyrcanian Forests by the Use of Sentinel-2 Data

by Fardin Moradi, Ali Asghar Darvishsefat, Manizheh Rajab Pourrahmati, Azade Deljouei and Stelian Alexandru Borz

Forests 2022, 13(1), 104; https://doi.org/10.3390/f13010104 - 12 Jan 2022

Cited by 27 | Viewed by 3344

Abstract

Due to the challenges brought by field measurements to estimate the aboveground biomass (AGB), such as the remote locations and difficulties in walking in these areas, more accurate and cost-effective methods are required, by the use of remote sensing. In this study, Sentinel-2 data were used for estimating the AGB in pure stands of Carpinus betulus (L., common hornbeam) located in the Hyrcanian forests, northern Iran. For this purpose, the diameter at breast height (DBH) of all trees thicker than 7.5 cm was measured in 55 square plots (45 × 45 m). In situ AGB was estimated using a local volume table and the specific density of wood. To estimate the AGB from remotely sensed data, parametric and nonparametric methods, including Multiple Regression (MR), Artificial Neural Network (ANN), k-Nearest Neighbor (kNN), and Random Forest (RF), were applied to a single image of the Sentinel-2, having as a reference the estimations produced by in situ measurements and their corresponding spectral values of the original spectral (B2, B3, B4, B5, B6, B7, B8, B8a, B11, and B12) and derived synthetic (IPVI, IRECI, GEMI, GNDVI, NDVI, DVI, PSSRA, and RVI) bands. Band 6 located in the red-edge region (0.740 nm) showed the highest correlation with AGB (r = −0.723). A comparison of the machine learning methods indicated that the ANN algorithm returned the best ABG-estimating performance (%RMSE = 19.9). This study demonstrates that simple vegetation indices extracted from Sentinel-2 multispectral imagery can provide good results in the AGB estimation of C. betulus trees of the Hyrcanian forests. The approach used in this study may be extended to similar areas located in temperate forests. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

23 pages, 5990 KiB

Open AccessArticle

Analysis of Factors Influencing Forest Loss in South Korea: Statistical Models and Machine-Learning Model

by Jeongmook Park, Byeoungmin Lim and Jungsoo Lee

Forests 2021, 12(12), 1636; https://doi.org/10.3390/f12121636 - 25 Nov 2021

Cited by 6 | Viewed by 2324

Abstract

Analyzing the current status of forest loss and its causes is crucial for understanding and preparing for future forest changes and the spatial pattern of forest loss. We investigated spatial patterns of forest loss in South Korea and assessed the effects of various factors on forest loss based on spatial heterogeneity. We used the local Moran’s I to classify forest loss spatial patterns as high–high clusters, low–low clusters, high–low outliers, and high–low outliers. Additionally, to assess the effect of factors on forest loss, two statistical models (i.e., ordinary least squares regression (OLS) and geographically weighted regression (GWR) models) and one machine-learning model (i.e., random forest (RF) model) were used. The accuracy of each model was determined using the R², RMSE, MAE, and AICc. Across South Korea, the forest loss rate was highest in the Seoul–Incheon–Gyeonggi region. Moreover, high–high spatial clusters were found in the Seoul–Incheon–Gyeonggi and Daejeon–Chungnam regions. Among the models, the GWR model was the most accurate. Notably, according to the GWR model, the main factors driving forest loss were road density, cropland area, number of households, and number of tertiary industry establishments. However, the factors driving forest loss had varying degrees of influence depending on the location. Therefore, our findings suggest that spatial heterogeneity should be considered when developing policies to reduce forest loss. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

14 pages, 4780 KiB

Open AccessArticle

Testing Forestry Digital Twinning Workflow Based on Mobile LiDAR Scanner and AI Platform

by Mihai Daniel Niță

Forests 2021, 12(11), 1576; https://doi.org/10.3390/f12111576 - 16 Nov 2021

Cited by 16 | Viewed by 3194

Abstract

Climate-smart forestry is a sustainable forest management approach for increasing positive climate impacts on society. As climate-smart forestry is focusing on more sustainable solutions that are resource-efficient and circular, digitalization plays an important role in its implementation. The article aimed to validate an automatic workflow of processing 3D pointclouds to produce digital twins for every tree on large 1-ha sample plots using a GeoSLAM mobile LiDAR scanner and VirtSilv AI platform. Specific objectives were to test the efficiency of segmentation technique developed in the platform for individual trees from an initial cloud of 3D points observed in the field and to quantify the efficiency of digital twinning by comparing the automatically generated results of (DBH, H, and Volume) with traditional measurements. A number of 1399 trees were scanned with LiDAR to create digital twins and, for validation, were measured with traditional tools such as forest tape and vertex. The segmentation algorithm developed in the platform to extract individual 3D trees recorded an accuracy varying between 95 and 98%. This result was higher in accuracy than reported by other solutions. When compared to traditional measurements the bias for diameter at breast height (DBH) and height was not significant. Digital twinning offers a blockchain solution for digitalization, and AI platforms are able to provide technological advantage in preserving and restoring biodiversity with sustainable forest management. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

11 pages, 2372 KiB

Open AccessFeature PaperArticle

Monitoring of Carriageway Cross Section Profiles on Forest Roads: Assessment of an Ultrasound Data Based Road Scanner with TLS Data Reference

by Michael Starke, Anton Kunneke and Martin Ziesak

Forests 2021, 12(9), 1191; https://doi.org/10.3390/f12091191 - 02 Sep 2021

Cited by 3 | Viewed by 1865

Abstract

Forest roads are an important element in forest management as they provide infrastructure for different forest stakeholder groups. Over time, a variety of road assessment concepts for better planning were initiated. The monitoring of the surface cross-section profile of forest roads particularly offers the possibility to take early action in restoring a road segment and avoiding higher future costs. One vehicle-based monitoring system that relies on ultrasound sensors addresses this topic. With advantages in its dirt influence tolerance and high temporal resolution, but shortcomings in horizontal and vertical measuring accuracy, the system was tested against high resolution terrestrial laser scanner (TLS) data to find and assess working scenarios that fit the low- resolution measuring principle. In a related field test, we found low correct road geometry interpretation rates of 54.3% but rising to 91.2% under distinctive geometric properties. The further applied line- and segment-based method used to transform the TLS data to fit the road scanner measuring method allows the transfer of the road scanner evaluation principle to point-cloud or raster data of different origins. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

23 pages, 5734 KiB

Open AccessArticle

A Collaborative Region Detection and Grading Framework for Forest Fire Smoke Using Weakly Supervised Fine Segmentation and Lightweight Faster-RCNN

by Jin Pan, Xiaoming Ou and Liang Xu

Forests 2021, 12(6), 768; https://doi.org/10.3390/f12060768 - 10 Jun 2021

Cited by 29 | Viewed by 3030

Abstract

Forest fires are serious disasters that affect countries all over the world. With the progress of image processing, numerous image-based surveillance systems for fires have been installed in forests. The rapid and accurate detection and grading of fire smoke can provide useful information, which helps humans to quickly control and reduce forest losses. Currently, convolutional neural networks (CNN) have yielded excellent performance in image recognition. Previous studies mostly paid attention to CNN-based image classification for fire detection. However, the research of CNN-based region detection and grading of fire is extremely scarce due to a challenging task which locates and segments fire regions using image-level annotations instead of inaccessible pixel-level labels. This paper presents a novel collaborative region detection and grading framework for fire smoke using a weakly supervised fine segmentation and a lightweight Faster R-CNN. The multi-task framework can simultaneously implement the early-stage alarm, region detection, classification, and grading of fire smoke. To provide an accurate segmentation on image-level, we propose the weakly supervised fine segmentation method, which consists of a segmentation network and a decision network. We aggregate image-level information, instead of expensive pixel-level labels, from all training images into the segmentation network, which simultaneously locates and segments fire smoke regions. To train the segmentation network using only image-level annotations, we propose a two-stage weakly supervised learning strategy, in which a novel weakly supervised loss is proposed to roughly detect the region of fire smoke, and a new region-refining segmentation algorithm is further used to accurately identify this region. The decision network incorporating a residual spatial attention module is utilized to predict the category of forest fire smoke. To reduce the complexity of the Faster R-CNN, we first introduced a knowledge distillation technique to compress the structure of this model. To grade forest fire smoke, we used a 3-input/1-output fuzzy system to evaluate the severity level. We evaluated the proposed approach using a developed fire smoke dataset, which included five different scenes varying by the fire smoke level. The proposed method exhibited competitive performance compared to state-of-the-art methods. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

26 pages, 15217 KiB

Open AccessArticle

Development of a Modality-Invariant Multi-Layer Perceptron to Predict Operational Events in Motor-Manual Willow Felling Operations

by Stelian Alexandru Borz

Forests 2021, 12(4), 406; https://doi.org/10.3390/f12040406 - 29 Mar 2021

Cited by 7 | Viewed by 2313

Abstract

Motor-manual operations are commonly implemented in the traditional and short rotation forestry. Deep knowledge of their performance is needed for various strategic, tactical and operational decisions that rely on large amounts of data. To overcome the limitations of traditional analytical methods, Artificial Intelligence (AI) has been lately used to deal with various types of signals and problems to be solved. However, the reliability of AI models depends largely on the quality of the signals and on the sensing modalities used. Multimodal sensing was found to be suitable in developing AI models able to learn time and location-related data dependencies. For many reasons, such as the uncertainty of preserving the sensing location and the inter- and intra-variability of operational conditions and work behavior, the approach is particularly useful for monitoring motor-manual operations. The main aim of this study was to check if the use of acceleration data sensed at two locations on a brush cutter could provide a robust AI model characterized by invariance to data sensing location. As such, a Multi-Layer Perceptron (MLP) with backpropagation was developed and used to learn and classify operational events from bimodally-collected acceleration data. The data needed for training and testing was collected in the central part of Romania. Data collection modalities were treated by fusion in the training dataset, then four single-modality testing datasets were used to check the performance of the model on a binary classification problem. Fine tuning of the regularization parameters (α term) has led to acceptable testing and generalization errors of the model measured as the binary cross-entropy (log loss). Irrespective of the hyperparameters’ tunning strategy, the classification accuracy (CA) was found to be very high, in many cases approaching 100%. However, the best models were those characterized by α set at 0.0001 and 0.1, for which the CA in the test datasets ranged from 99.1% to 99.9% and from 99.5% to 99.9%, respectively. Hence, data fusion in the training set was found to be a good strategy to build a robust model, able to deal with data collected by single modalities. As such, the developed MLP model not only removes the problem of sensor placement in such applications, but also automatically classifies the events in the time domain, enabling the integration of data collection, handling and analysis in a simple less resource-demanding workflow, and making it a feasible alternative to the traditional approach to the problem. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Figure 1

17 pages, 14033 KiB

Open AccessArticle

A Forest Fire Detection System Based on Ensemble Learning

by Renjie Xu, Haifeng Lin, Kangjie Lu, Lin Cao and Yunfei Liu

Forests 2021, 12(2), 217; https://doi.org/10.3390/f12020217 - 13 Feb 2021

Cited by 317 | Viewed by 18261

Abstract

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. Full article

(This article belongs to the Special Issue Electronics, Close-Range Sensors and Artificial Intelligence in Forestry)

► Show Figures

Journal Menu

Journal Browser

Electronics, Close-Range Sensors and Artificial Intelligence in Forestry

Share This Special Issue

Special Issue Editors

Special Issue Information

Keywords

Published Papers (13 papers)

Editorial

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI