High-resolution Unmanned Aerial Vehicles Images Research Articles

Classifying Stand Compositions in Clover Grass Based on High-Resolution Multispectral UAV Images

In organic farming, clover is an important basis for green manure in crop rotation systems due to its nitrogen-fixing effect. However, clover is often sown in mixtures with grass to achieve a yield-increasing effect. In order to determine the quantity and distribution of clover and its influence on the subsequent crops, clover plants must be identified at the individual plant level and spatially differentiated from grass plants. In practice, this is usually done by visual estimation or extensive field sampling. High-resolution unmanned aerial vehicles (UAVs) offer a more efficient alternative. In the present study, clover and grass plants were classified based on spectral information from high-resolution UAV multispectral images and texture features using a random forest classifier. Three different timestamps were observed in order to depict the phenological development of clover and grass distributions. To reduce data redundancy and processing time, relevant texture features were selected based on a wrapper analysis and combined with the original bands. Including these texture features, a significant improvement in classification accuracy of up to 8% was achieved compared to a classification based on the original bands only. Depending on the phenological stage observed, this resulted in overall accuracies between 86% and 91%. Subsequently, high-resolution UAV imagery data allow for precise management recommendations for precision agriculture with site-specific fertilization measures.

Application of an improved U-Net with image-to-image translation and transfer learning in peach orchard segmentation

Peach cultivation holds a significant economic importance, and obtaining the spatial distribution of peach orchards is helpful for yield prediction and precision agriculture. In this study, we introduce a new U-Net semantic segmentation model, utilizing ResNet50 as a backbone network, augmented with an Efficient Multi-Scale Attention (EMA) mechanism module and a LayerScale adaptive scaling parameter. To address style differences between images from Unmanned Aerial Vehicle (UAV), Google Earth, and Sentinel-2 satellite, we incorporate Cycle-Consistent Generative Adversarial Networks (CycleGAN). This synthesis ensures that UAV images conform to a comparable style found in Google Earth and Sentinel-2 images, while feature details of high spatial resolution UAV images are transferred to Google Earth and Sentinel-2 images through transfer learning. The results demonstrate that using ResNet50 as a backbone network for the U-Net model yields higher accuracy compared to using VGG16 for the U-Net model. Specifically, the Mean Intersection over Union (MIoU) values for UAV and Sentinel-2 images are higher by 0.49 % and 0.95 %, respectively. The MIoU values for UAV, Google Earth, and Sentinel-2 images increased by 0.87 %, 1.71 %, and 1.74 %, respectively, with the introduction of EMA. Additionally, with the introduction of LayerScale adaptive scaling parameters, the MIoU values increased by 0.31 %, 0.33 %, and 1.44 %, respectively, further enhancing the segmentation accuracy of the model. After applying CycleGAN and transfer learning, the MIoU increased by 1.02 %, 0.15 %, and 1.57 % for UAV, Google Earth, and Sentinel-2 images, respectively, resulting in MIoU values of 97.39 %, 92.08 %, and 84.54 %. The comparative analysis with DeepLabV3+, PSPNet, and HRNet models demonstrates the superior mapping performance of the proposed method. Moreover, the method exhibits good generalization and mapping speed across six test sites in the research area. Overall, this approach ensures high precision and efficiency in peach orchard mapping, accommodating various spatial resolutions, and holds potential for addressing diverse requirements in peach orchard mapping applications.

Evapotranspiration partitioning by integrating eddy covariance, micro-lysimeter and unmanned aerial vehicle observations: A case study in the North China Plain

Partitioning of evapotranspiration (ET) into soil evaporation (E) and plant transpiration (T) poses a significant challenge. This study proposed a novel approach that combines eddy covariance, micro-lysimeter, and high-resolution unmanned aerial vehicle (UAV) images within the flux footprint to optimize the canopy and soil resistances of the Shuttleworth-Wallace model (S-W) using a multi-objective optimization scheme. A two-year experiment was conducted at the Yucheng and Luancheng sites to investigate the effectiveness of the proposed method for the summer maize-wheat winter system in the North China Plain. Our results showed that the S-W model with multi-objective optimization by using high-resolution UAV images within the flux footprint significantly improved the estimation accuracy of ET components compared with the traditional one- and two-objective optimization schemes. The estimation of ET, T, and E of our method had an average root mean square error of 0.67 mm day−1, 0.66 mm day−1, and 0.28 mm day−1, respectively. The optimized S-W model using high-resolution UAV data within the flux footprint produced better ET partitioning than that optimized using Moderate Resolution Imaging Spectroradiometer (MODIS) data. We then applied the proposed method to estimate ET and ET components in the North China Plain. This study highlighted the importance of utilizing multi-source data (especially high-resolution UAV images) and multi-objective optimization for accurate ET partitioning.

Utilizing a single-temporal full polarimetric Gaofen-3 SAR image to map coseismic landslide inventory following the 2017 Mw 7.0 Jiuzhaigou earthquake (China)

On August 8, 2017, a magnitude 7.0 earthquake struck Jiuzhaigou County in Sichuan Province, triggering numerous coseismic landslides. The prompt identification of these landslides is imperative for emergency rescue efforts and post-earthquake hazard assessments. Optical satellite and unmanned aerial vehicle (UAV) images are often obstructed by cloud cover and fog following earthquakes. In contrast, polarimetric synthetic aperture radar (PolSAR), unaffected by adverse weather conditions, emerges as an indispensable tool. However, the utilization of spaceborne single-temporal SAR for mapping the inventory of coseismic landslides is infrequent and encounters constraints due to several limitations. In this study, we analyzed the amplitude feature and polarimetric decomposition of multiple ground categories in a full PolSAR image, and proposed an automated method to accurately identify coseismic landslides using a single-temporal full PolSAR image. The coseismic landslide inventory following the Jiuzhaigou Earthquake was mapped and validated using high-resolution UAV images. Detailed analysis was conducted to identify error sources leading to omissions and false positives. Additionally, we evaluated various machine learning models to compare their performance with our proposed method. Finally, we conducted a comprehensive discussion on the strengths and weaknesses of different data types (PolSAR, optical satellite, and UAV) for coseismic landslide identification. Our results indicate that the proposed PolSAR-based method achieves high accuracy in coseismic landslide inventory mapping, offering an effective solution for timely post-earthquake emergency responses in complex environments and all weather conditions in the future.

EROSION RATE ESTIMATION IN KUALA TERENGGANU COASTLINE FOR SUSTAINABLE COASTAL COMMUNITY

Kuala Terengganu is located on the east coast of Peninsular Malaysia and its coastline has been attracting various economic activities such as tourism, agriculture, and food industries. However, studies regarding erosion management for the populated sustainable community in Terengganu remain lacking. The purpose of this study was to quantify coastal erosion using geospatial and statistical approaches. It involved the utilisation of high-resolution SPOT-5 satellite imagery and unmanned aerial vehicles (UAVs) to monitor the coastline changes along Kampung Batu Rakit to Pantai Tok Jembal in Kuala Terengganu. This study proves that the geospatial technique based on high-resolution UAV and SPOT-5 images is suitable for the determination and analysis of coastline erosion.

Monitoring Individual Tree Phenology in a Multi-Species Forest Using High Resolution UAV Images

Monitoring tree phenology is important for understanding ecosystem functioning and for assessing ecosystem responses to climate change. Satellite imagery offers open-access global coverage but is restricted to forest-level analyses, due to its coarse spatial resolution. Unmanned aerial vehicle (UAV) imagery can monitor phenology at the individual tree level by utilizing a centimeter-scale resolution. Two research objectives were identified for this study: (1) to derive phenological metrics at the individual tree level, using various vegetation indices (VIs); and (2) to assess the accuracy of automatic crown delineation in a diverse ecosystem. To achieve this, fourteen multi-spectral UAV flights were performed, and the ability of the normalized difference vegetation index (NDVI), enhanced vegetation index 2 (EVI2), optimized soil-adjusted vegetation index (OSAVI), and chlorophyll index red-edge (CIre) to model seasonal phenology was assessed. A double logistic model was fitted on the VI observations for each individual tree, to derive the start of season (SOS) and end of season (EOS). Individual tree crowns were delineated automatically using marker-controlled watershed segmentation (MCWS), and the treetops were identified using a local maximum filter (LMF). Overall, the automatic segmentation performed well (F-score: 0.79, IoU: 0.58), with higher accuracies in single-species areas, while it underperformed in complex mixed forest structures. All VIs captured a strong seasonal signal for the deciduous trees and derived SOS and EOS estimates consistent with literature and ground observations. General phenological patterns included an early silver birch SOS, a quick beech budburst, and large within-species phenology variations for oak trees. Seasonal VI variation for coniferous evergreen trees was limited, and the resulting phenology estimates proved unreliable. In conclusion, these findings emphasize the capabilities of UAV imagery for individual tree crown phenology monitoring. However, they also show the difficulty of monitoring evergreen phenology with the commonly-used VIs and stress the need for further investigations.

Classification of Rice Growth Stage on UAV Image Based on Convolutional Neural Network Method

Currently, the majority of the agricultural sector in Indonesia is carried out by small communities. Half of the Indonesian people (approximately 10 million people) work in the agricultural sector and utilize agricultural land. Some of the tools used by farmers are still using traditional tools, but some are already using modern farming tools. In general, agricultural tools are divided into 3 categories, namely agricultural tools used before the seeds are planted, agricultural tools used when caring for seedlings that are growing and developing, and agricultural tools used when harvesting. One of the technologies used in agriculture is the use of drones or Unmanned Aerial Vehicles (UAV) in the process of sowing fertilizers and seeds and spraying pesticides. The current use of UAVs supports agriculture with manual operation and based on GPS waypoint positioning. In the process, the visual aspects that can be obtained from the UAV have not been considered, so the treatment carried out on agricultural land is the same. The problem of similarity in treatment can lead to similar treatment on heterogeneous agricultural land. Agricultural land should be treated according to the conditions of the land. Because the condition of the land will affect the growth of the planted vegetation. Another problem found in agricultural land is the different rice growth in each paddy field. Rice growth can be seen by farmers through visual aspects but farmers cannot directly see the visual condition of rice growth as a whole because of the large area of land. Utilization of UAV by taking high-resolution aerial imagery can provide visuals of the overall condition of rice from various angles of image capture. The general objective of this research is to classify rice growth on high resolution UAV images based on the Convolutional Neural Network (CNN). The data used in this study were acquired using a multirotor UAV in the same rice field area. The data consists of 500 images consisting of 5 groups. Group 1-2 is the vegetative phase, group 3 is the generative phase and group 4-5 is the ripening phase. CNN is used to conduct training with variations of epochs are 100, 250 and 500. The best accuracy results are obtained in the training epoch 500 with 96% of Accuration

Car Detection from Very High-Resolution UAV Images Using Deep Learning Algorithms

It is important to determine car density in parking lots, especially in hospitals, large enterprises, and residential areas, which are used intensively, in terms of executing existing management systems and making precise plans for the future. In this study, cars in parking lots were detected using high-resolution unmanned aerial vehicle (UAV) images with deep learning methods. We tested the performance of the two approaches by determining the number of cars in a parking lot using the You Only Look Once (YOLOv3) and Mask Region–Based Convolutional Neural Networks (Mask R-CNN) approaches as deep learning methods and the deep learning tool of Esri ArcGIS Pro. High-resolution UAV images were processed by photogrammetry and used as input products for the R-CNN and YOLOv3 algorithm. Recall, F1 score, precision ratio/uncertainty accuracy, and average producer accuracy of products automatically extracted with the algorithm were determined as 0.862/0.941, 0.874/0.946, 0.885/0.951, and 0.776/0.897 for R-CNN and YOLOv3, respectively.

Quantitative Assessment Framework for Non-Structural Bird’s Nest Risk Information of Transmission Tower in High-Resolution UAV Images

As the main cause of foreign body intrusion, bird’s nests pose a significant potential risk to high-voltage transmission towers. Currently, most research uses convolutional neural networks (CNNs) to detect partial images of towers captured by manually operated unmanned aerial vehicles (UAVs). However, these methods have difficulties in learning the locations and potential risk information of bird’s nests and generating structured data results. To address these challenges, this paper presents a quantitative assessment framework for non-structural bird’s nest risk information of transmission towers using high-resolution UAV panoramic images. The integrity of the tower’s structure information is maintained by using panoramic UAV images. The proposed method involves estimating the tower’s pose and determining whether each key point of the tower has a bird’s nest, transforming the non-structural bird’s nest locations and risk information into structured data. A risk score is used to classify the level of potential risk caused by bird’s nests. The experiments demonstrate that the proposed framework can accurately detect the key points of five types of towers, with an F1-score of 83.09% for bird’s nest detection and an accuracy of 91.59% for risk assessment.

Thermal UAV Image Super-Resolution Guided by Multiple Visible Cues

Unmanned aerial vehicle (UAV) thermal-imaging has received much attention, but the insufficient image resolution caused by thermal imaging systems is still a crucial problem that limits the understanding of thermal UAV images. However, high-resolution visible images are relatively easy to access, and it is thus valuable for exploring useful information from visible image to assist thermal UAV image super-resolution (SR). In this article, we propose a novel multiconditioned guidance network (MGNet) to effectively mine the information of visible images for thermal UAV image SR. High-resolution visible UAV images usually contain salient appearance, semantic, and edge information, which plays a critical role in boosting the performance of thermal UAV image SR. Therefore, we design an effective multicue guidance module (MGM) to leverage the appearance, edge, and semantic cues from visible images to guide thermal UAV image SR. In addition, we build the first benchmark dataset for the task of thermal UAV image SR guided by visible images. It is collected by a multimodal UAV platform and composes of 1025 pairs of manually aligned visible and thermal images. Extensive experiments on the built dataset show that our MGNet can effectively leverage useful information from visible images to improve the performance of thermal UAV image SR and perform well against several state-of-the-art methods. The dataset is available at: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/mmic-lcl/Datasets-and-benchmark-code</uri> .

Recognition of Earthquake Surface Ruptures Using Deep Learning

Investigating post-earthquake surface ruptures is important for understanding the tectonics of seismogenic faults. The use of unmanned aerial vehicle (UAV) images to identify post-earthquake surface ruptures has the advantages of low cost, fast data acquisition, and high data processing efficiency. With the rapid development of deep learning in recent years, researchers have begun using it for image crack detection. However, due to the complex background and diverse characteristics of the surface ruptures, it remains challenging to quickly train an effective automatic earthquake surface rupture recognition model on a limited number of samples. This study proposes a workflow that applies an image segmentation algorithm based on convolutional neural networks (CNNs) to extract cracks from post-earthquake UAV images. We selected the 16-layer visual geometry group (VGG16) network as the primary network architecture. Then, we improved the VGG16 network and deleted several convolutional layers to reduce computation and memory consumption. Moreover, we added dilated convolution and atrous spatial pyramid pooling (ASPP) to make the network perform well in the surface crack identification of post-earthquake UAV images. We trained the proposed method using the data of the MS 7.4 Maduo earthquake and obtained a model that could automatically identify and draw small and irregular surface ruptures from high-resolution UAV images.

Geospatial-based machine learning techniques for land use and land cover mapping using a high-resolution unmanned aerial vehicle image

Bangladesh is primarily an agricultural country where technological advancement in the agricultural sector can ensure the acceleration of economic growth and ensure long-term food security. This research was conducted in the south-western coastal zone of Bangladesh, where rice is the main crop and other crops are also grown. Land use and land cover (LULC) classification using remote sensing techniques such as the use of satellite or unmanned aerial vehicle (UAV) images can forecast the crop yield and can also provide information on weeds, nutrient deficiencies, diseases, etc. to monitor and treat the crops. Depending on the reflectance received by sensors, remotely sensed images store a digital number (DN) for each pixel. Traditionally, these pixel values have been used to separate clusters and classify various objects. However, it frequently generates a lot of discontinuity in a particular land cover, resulting in small objects within a land cover that provide poor image classification output. It is called the salt-and-pepper effect. In order to classify land cover based on texture, shape, and neighbors, Pixel-Based Image Analysis (PBIA) and Object-Based Image Analysis (OBIA) methods use digital image classification algorithms like Maximum Likelihood (ML), K-Nearest Neighbors (KNN), k-means clustering algorithm, etc. to smooth this discontinuity. The authors evaluated the accuracy of both the PBIA and OBIA approaches by classifying the land cover of an agricultural field, taking into consideration the development of UAV technology and enhanced image resolution. For classifying multispectral UAV images, we used the KNN machine learning algorithm for object-based supervised image classification and Maximum Likelihood (ML) classification (parametric) for pixel-based supervised image classification. Whereas, for unsupervised classification using pixels, we used the K-means clustering technique. For image analysis, Near-infrared (NIR), Red (R), Green (G), and Blue (B) bands of a high-resolution ground sampling distance (GSD) 0.0125m UAV image was used in this research work. The study found that OBIA was 21% more accurate than PBIA, indicating 94.9% overall accuracy. In terms of Kappa statistics, OBIA was 27% more accurate than PBIA, indicating Kappa statistics accuracy of 93.4%. It indicates that OBIA provides better classification performance when compared to PBIA for the classification of high-resolution UAV images. This study found that by suggesting OBIA for more accurate identification of types of crops and land cover, which will help crop management, agricultural monitoring, and crop yield forecasting be more effective.

Estimating canopy-scale chlorophyll content in apple orchards using a 3D radiative transfer model and UAV multispectral imagery

Chlorophyll content is a key trait for understanding the functioning of agroforestry ecosystems and has important implications for leaf and canopy photosynthesis. However, fine-scale monitoring of canopy chlorophyll content (CCC) of individual fruit trees is rather challenging. This study aims to use a 3D radiative transfer model (RTM) and proposes a joint inversion model based on prior knowledge to estimate the CCC of individual tree crowns (ITCs) in apple orchards. The widely recognized 3D RTM LESS (large-scale remote sensing data and image simulation framework over heterogeneous 3D scenes) was adopted for large-scale apple orchard 3D scenes radiative transfer computing and image simulation. LESS was first evaluated with unmanned aerial vehicle (UAV) multispectral imagery and the results showed that it reasonably characterized the reflectance of apple tree canopies (RMSE = 0.02). An original look-up table (LUT) with reflectance was then produced using LESS, and the final vegetation indices LUT (VI LUT) including Normalized Difference Vegetation Index (NDVI), Green Chlorophyll Index (CIgreen), Red edge Chlorophyll Index (CIred edge) and Green NDVI (GNDVI) was generated from the original LUT form VI interpolation. A physically-based joint inversion model coupling prior knowledge of leaf pigments and leaf area index (LAI) was developed to estimate the CCC of ITCs from high-resolution UAV images. The solution first used linear interpolation to produce a weighted VI LUT corresponding to the sample based on estimated LAI. Linear interpolation was then adopted to screen multiple combinations of leaf chlorophylla+b (Cab) and leaf carotenoids (Cxc) contents from the VI LUT. A prior relationship between Cab and Cxc was finally used to regularize the constraints on multiple VI combinations and determine the estimation of Cab and CCC. The joint inversion model demonstrated an accurate estimation of CCC of ITCs. The model driven by GNDVI yielded the highest result for CCC estimation (R2 = 0.84, RMSE = 24.12 μg/cm2). In addition, CIgreen (R2 = 0.82, RMSE = 32.22 μg/cm2) and CIred edge (R2 = 0.81, RMSE = 34.05 μg/cm2) also achieved satisfactory results. The proposed model facilitates CCC estimation of ITCs from high-resolution imagery in heterogeneous orchard canopies, which is important for advancing the precise nutrition management of fruit trees.

ARSD: An Adaptive Region Selection Object Detection Framework for UAV Images

Due to the rapid development of deep learning, the performance of object detection has greatly improved. However, object detection in high-resolution Unmanned Aerial Vehicles images remains a challenging problem for three main reasons: (1) the objects in aerial images have different scales and are usually small; (2) the images are high-resolution but state-of-the-art object detection networks are of a fixed size; (3) the objects are not evenly distributed in aerial images. To this end, we propose a two-stage Adaptive Region Selection Detection framework in this paper. An Overall Region Detection Network is first applied to coarsely localize the object. A fixed points density-based targets clustering algorithm and an adaptive selection algorithm are then designed to select object-dense sub-regions. The object-dense sub-regions are sent to a Key Regions Detection Network where results are fused with the results at the first stage. Extensive experiments and comprehensive evaluations on the VisDrone2021-DET benchmark datasets demonstrate the effectiveness and adaptiveness of the proposed framework. Experimental results show that the proposed framework outperforms, in terms of mean average precision (mAP), the existing baseline methods by 2.1% without additional time consumption.

TAU: A framework for video-based traffic analytics leveraging artificial intelligence and unmanned aerial systems

Smart traffic engineering and intelligent transportation services are in increasing demand from governmental authorities to optimize traffic performance and thus reduce energy costs, increase the drivers’ safety and comfort, ensure traffic laws enforcement, and detect traffic violations. In this paper, we address this challenge, and we leverage the use of Artificial Intelligence (AI) and Unmanned Aerial Vehicles (UAVs) to develop an AI-integrated video analytics framework, called TAU (Traffic Analysis from UAVs), for automated traffic analytics and understanding. Unlike previous works on traffic video analytics, we propose an automated object detection and tracking pipeline from video processing to advanced traffic understanding using high-resolution UAV images. TAU combines six main contributions. First, it proposes a pre-processing algorithm to adapt the high-resolution UAV image as input to the object detector without lowering the resolution. This ensures an excellent detection accuracy from high-quality features, particularly the small size of detected objects from UAV images. Second, it introduces an algorithm for recalibrating the vehicle coordinates to ensure that vehicles are uniquely identified and tracked across the multiple crops of the same frame. Third, it presents a speed calculation algorithm based on accumulating information from successive frames. Fourth, TAU counts the number of vehicles per traffic zone based on the Ray Tracing algorithm. Fifth, TAU has a fully independent algorithm for crossroad arbitration based on the data gathered from the different zones surrounding it. Sixth, TAU introduces a set of algorithms for extracting twenty-four types of insights from the raw data collected. These insights facilitate the traffic understanding using curves, histograms, heatmaps, and animations. The present work presents a valuable added value for academic researchers and transportation engineers to automate the traffic video analytics process and extract useful insights to optimize traffic performance. TAU is a ready-to-use framework for any Transportation Engineer to better understand and manage daily road traffic (video demonstrations are provided here: https://youtu.be/wXJV0H7LviU and here: https://youtu.be/kGv0gmtVEbI). The source code is available at: https://github.com/bilel-bj/TAU.

Combining Spectral and Textural Information from UAV RGB Images for Leaf Area Index Monitoring in Kiwifruit Orchard

The use of a fast and accurate unmanned aerial vehicle (UAV) digital camera platform to estimate leaf area index (LAI) of kiwifruit orchard is of great significance for growth, yield estimation, and field management. LAI, as an ideal parameter for estimating vegetation growth, plays a significant role in reflecting crop physiological process and ecosystem function. At present, LAI estimation mainly focuses on winter wheat, corn, soybean, and other food crops; in addition, LAI on forest research is also predominant, but there are few studies on the application of orchards such as kiwifruit. Concerning this study, high-resolution UAV images of three growth stages of kiwifruit orchard were acquired from May to July 2021. The extracted significantly correlated spectral and textural parameters were used to construct univariate and multivariate regression models with LAI measured for corresponding growth stages. The optimal model was selected for LAI estimation and mapping by comparing the stepwise regression (SWR) and random forest regression (RFR). Results showed the model combining texture features was superior to that only based on spectral indices for the prediction accuracy of the modeling set, with the R2 of 0.947 and 0.765, RMSE of 0.048 and 0.102, and nRMSE of 7.99% and 16.81%, respectively. Moreover, the RFR model (R2 = 0.972, RMSE = 0.035, nRMSE = 5.80%) exhibited the best accuracy in estimating LAI, followed by the SWR model (R2 = 0.765, RMSE = 0.102, nRMSE = 16.81%) and univariate linear regression model (R2 = 0.736, RMSE = 0.108, nRMSE = 17.84%). It was concluded that the estimation method based on UAV spectral parameters combined with texture features can provide an effective method for kiwifruit growth process monitoring. It is expected to provide scientific guidance and practical methods for the kiwifruit management in the field for low-cost UAV remote sensing technology to realize large area and high-quality monitoring of kiwifruit growth, thus providing a theoretical basis for kiwifruit growth investigation.

Color Balance Method of Dense Point Cloud in Landslides Area Based on UAV Images

Landslides are a natural geological phenomenon and often cause economic losses, property damage, and loss of lives. The detection and recognition based on the dense point cloud derived from Unmanned Aerial Vehicle (UAVs) images play an important role in understanding the mechanism of landslides. However, the brightness difference of UAV images with large altitude differences usually makes the dense point cloud within a landslide area show color stratification, and the processing results of popular commercial and open-source software are not ideal. In this study, a multi-view UAV image brightness compensation method based on regular 3D points is proposed, and is combined with the open-source software OpenMVG and OpenMVS to obtain a dense point cloud with balanced color on the landslide’s surface. The selected study area is located in the active landslide region caused by the Tengqing coal mining area in Guizhou Province, China. The terrain following flight method was used to acquire complete high resolution UAV images of the study area. Experimental results show that the global gain and offset compensation methods proposed in this study can effectively reduce the 22.3% brightness difference in the overlapping area of the adjacent images. The comparison results and the processing results of other commercial software show that the method can be applied to the color balance of the dense point cloud for UAV images with large-altitude differences.

Transformer Neural Network for Weed and Crop Classification of High Resolution UAV Images

Monitoring crops and weeds is a major challenge in agriculture and food production today. Weeds compete directly with crops for moisture, nutrients, and sunlight. They therefore have a significant negative impact on crop yield if not sufficiently controlled. Weed detection and mapping is an essential step in weed control. Many existing research studies recognize the importance of remote sensing systems and machine learning algorithms in weed management. Deep learning approaches have shown good performance in many agriculture-related remote sensing tasks, such as plant classification, disease detection, etc. However, despite the success of these approaches, they still face many challenges such as high computation cost, the need of large labelled datasets, intra-class discrimination (in growing phase weeds and crops share many attributes similarity as color, texture, and shape), etc. This paper aims to show that the attention-based deep network is a promising approach to address the forementioned problems, in the context of weeds and crops recognition with drone system. The specific objective of this study was to investigate visual transformers (ViT) and apply them to plant classification in Unmanned Aerial Vehicles (UAV) images. Data were collected using a high-resolution camera mounted on a UAV, which was deployed in beet, parsley and spinach fields. The acquired data were augmented to build larger dataset, since ViT requires large sample sets for better performance, we also adopted the transfer learning strategy. Experiments were set out to assess the effect of training and validation dataset size, as well as the effect of increasing the test set while reducing the training set. The results show that with a small labeled training dataset, the ViT models outperform state-of-the-art models such as EfficientNet and ResNet. The results of this study are promising and show the potential of ViT to be applied to a wide range of remote sensing image analysis tasks.

Unmanned aerial vehicles and deep learning for assessment of anthropogenic marine debris on beaches on an island in a semi-enclosed sea in Japan

The increasing prevalence of marine debris is a global problem, and urgent action for amelioration is needed. Identifying hotspots where marine debris accumulates will enable effective control; however, knowledge on the location of accumulation hotspots remains incomplete. In particular, marine debris accumulation on beaches is a concern. Surveys of beaches require intensive human effort, and survey methods are not standardized. If marine debris monitoring is conducted using a standardized method, data from different regions can be compared. With an unmanned aerial vehicle (UAV) and deep learning computational methods, monitoring a wide area at a low cost in a standardized way may be possible. In this study, we aimed to identify marine debris on beaches through deep learning using high-resolution UAV images by conducting a survey on Narugashima Island in the Seto Inland Sea of Japan. The flight altitude relative to the ground was set to 5 m, and images of a 0.81-ha area were obtained. Flight was conducted twice: before and after the beach cleaning. The combination of UAVs equipped with a zoom lens and operation at a low altitude allows for the acquisition of high resolution images of 1.1 mm/pixel. The training dataset (2970 images) was annotated by using VoTT, categorizing them into two classes: ‘anthropogenic marine debris’ and ‘natural objects.’ Using RetinaNet, marine debris was identified with an average sensitivity of 51% and a precision of 76%. In addition, the abundance and area of marine debris coverage were estimated. In this study, it was revealed that the combination of UAVs and deep learning enables the effective identification of marine debris. The effects of cleanup activities by citizens were able to be quantified. This method can widely be used to evaluate the effectiveness of citizen efforts toward beach cleaning and low-cost long-term monitoring.

BASNet: Burned Area Segmentation Network for Real-Time Detection of Damage Maps in Remote Sensing Images

Since remote sensing images of post-fire vegetation are characterized by high resolution, multiple interferences, and high similarities between the background and the target area, it is difficult for existing methods to detect and segment the burned area in these images with sufficient speed and accuracy. In this paper, we apply Salient Object Detection (SOD) to burned area segmentation, the first time this has been done, and propose an efficient burned area segmentation network (BASNet) to improve the performance of unmanned aerial vehicle (UAV) high-resolution image segmentation. BASNet comprises positioning module and refinement module. The positioning module efficiently extracts high-level semantic features and general contextual information via global average pooling layer and convolutional block to determine the coarse location of the salient region. The refinement module adopts the convolutional block attention module to effectively discriminate the spatial location of objects. In addition, to effectively combine edge information with spatial location information in the lower layer of the network and the high-level semantic information in the deeper layer, we design the residual fusion module to perform feature fusion by level to obtain the prediction results of the network. Extensive experiments on two UAV datasets collected from Chongli in China and Andong in South Korea, demonstrate that our proposed BASNet significantly outperforms state-of-the-art SOD methods quantitatively and qualitatively. BASNet also achieves a promising prediction speed for processing high-resolution UAV images, thus providing wide-ranging applicability in post-disaster monitoring and management.