Unmanned Aerial Vehicle Images Research Articles

Soybean seedling detection and counting from UAV images based on an improved YOLOv8 Network

Abstract. The utilization of unmanned aerial vehicle (UAV) for soybean seedling detection is an effective way to estimate soybean yield, which plays a crucial role in agricultural planning and decision-making. However, the soybean seedlings objects in the UAV image are small, in clusters, and occluded each other, which makes it very challenging to achieve accurate object detection and counting. To address these issues, we optimize the YOLOv8 model and propose a GAS-YOLOv8 network, aiming to enhance the detection accuracy for the task of soybean seedling detection based on UAV images. Firstly, a global attention mechanism (GAM) is incorporated into the neck module of YOLOv8, which reallocates weights and prioritizes global information to more effectively extract soybean seedling features. Secondly, the CIOU loss function is replaced with the SIOU loss, which includes an angle loss term to guide the regression of bounding boxes. Experimental results show that, on the soybean seedling dataset, the proposed GAS-YOLOv8 model achieves a 1.3% improvement in mAP@0.5 and a 6% enhancement in detection performance in dense seedling areas, when compared to the baseline model YOLOv8s.When compared to other object detection models (YOLOv5, Faster R-CNN, etc.), the GAS-YOLOv8 model similarly achieved the best detection performance. These results demonstrate the effectiveness of the GAS-YOLOv8 in detecting dense soybean seedlings, providing more accurate theoretical support for subsequent yield estimation.

FS_YOLOv8: A Deep Learning Network for Ground Fissures Instance Segmentation in UAV Images of the Coal Mining Area

Abstract. The ground fissures caused by coal mining have seriously affected the ecological environment of the land. Timely and accurate identification and landfill treatment of ground fissures can avoid secondary geological disasters in coal mine areas. At present, the fissure identification methods based on deep learning show excellent performance on roads and walls, etc. Nevertheless, the automatic and reliable segmentation of ground fissures in remote sensing images poses a challenge for deep learning networks, due to the diverse and complex texture information included in the mining ground fissures and background. To overcome these challenges, we propose an improved YOLOv8 instance segmentation network to automatically and efficiently segment the ground fissures in coal mining areas. In detail, a model called FS_YOLOv8 is proposed. The DSPP (Dynamic Snake convolutional Pyramid Pooling) module is incorporated into the FS_YOLOv8 model to establish a multi-scale dynamic snake convolution feature aggregation structure. This module replaces the conventional convolution found in the SPPF module of YOLOv8 and aims to enhance the model's ability to extract features related to fissures with tubular structures. Furthermore, the D-LKA (Deformable Large Kernel Attention) module is employed to autonomously collect fissure context information. To enhance the detection capability of challenging samples in remote sensing images with intricate background and fissure texture, we employ a Slide Loss function. Ultimately, the ground fissure dataset of unmanned aerial vehicle (UAV) images in coal mine areas is subjected to experimental analysis. The experimental findings demonstrate that FS_YOLOv8 exhibits exceptional proficiency in segmenting ground fissures within intricate and expansive mining areas.

An improved algorithm based on YOLOv5 for detecting Ambrosia trifida in UAV images

A YOLOv5-based YOLOv5-KE unmanned aerial vehicle (UAV) image detection algorithm is proposed to address the low detection accuracy caused by the small size, high density, and overlapping leaves of Ambrosia trifida targets in UAV images. The YOLOv5-KE algorithm builds upon the YOLOv5 algorithm by adding a micro-scale detection layer, adjusting the hierarchical detection settings based on k-Means for Anchor Box, improving the loss function of CIoU, reselecting and improving the detection box fusion algorithm. Comparative validation experiments of the YOLOv5-KE algorithm for Ambrosia trifida recognition were conducted using a self-built dataset. The experimental results show that the best detection accuracy of Ambrosia trifida in UAV images is 93.9%, which is 15.2% higher than the original YOLOv5. Furthermore, this algorithm also outperforms other existing object detection algorithms such as YOLOv7, DC-YOLOv8, YOLO-NAS, RT-DETR, Faster RCNN, SSD, and Retina Net. Therefore, YOLOv5-KE is a practical algorithm for detecting Ambrosia trifida under complex field conditions. This algorithm shows good potential in detecting weeds of small, high-density, and overlapping leafy targets in UAV images, it could provide technical reference for the detection of similar plants.

Updating Orthophotos around Gas Pipelines based on "Cloud Control" Photogrammetry

Abstract. As a clean energy source, natural gas is widely used and primarily transported through long-distance pipelines. Regular maintenance and inspection of long-distance gas pipelines are crucial tasks. Due to the extensive coverage and distance of these pipelines, the workload is enormous. It is necessary to first identify areas of change, which can be carried out using multiple sets of orthophotos produced by unmanned aerial vehicles (UAVs). However, UAV images have small footprints and significant geometric distortions, requiring a large number of ground control points (GCPs) for accurate positioning. Measuring these points in the field is challenging and time-consuming, becoming a key factor limiting the rapid production of orthophotos. To overcome this challenge, this paper introduces the "cloud control" photogrammetry technology to achieve fully automatic updates of orthophotos around long-distance pipelines, providing foundational data for the maintenance and inspection of these gas pipelines. This method replaces GCPs with images containing known orientation parameters, serving as control information. By matching tie points between new and old images, the "cloud control points" are transferred to the new images, enabling the image registration and production of orthophotos. The experiments conducted on the Fumin and Zhaotong segments of a long-distance gas pipeline in Yunnan Province demonstrate that, for UAV images with a ground resolution of 0.05 meters, using the "cloud control" method achieves a planar accuracy of 0.05 meters and an elevation accuracy of 0.07 meters. These results are comparable to the accuracy obtained by orienting the results using GCPs.

Accuracy Assessment of UAV Photogrammetry System with RTK Measurements for Direct Georeferencing

Abstract. The direct georeferencing accuracy of unmanned aerial vehicle (UAV) images with real-time kinematic (RTK) measurements is a concerned topic in the community of photogrammetry. This study assesses the capabilities of a multi-rotor platform equipped with RTK technology, specifically a DJI Phantom 4 RTK UAV, for robust direct georeferencing. The UAV surveyed a square and a building at Wuhan University to assess the accuracy and spatial consistency of direct georeferencing in close-range photography. We tested checkpoint errors under various ground control points (GCPs) configurations. The results show that without GCP, an analysis of 71 spatially distributed checkpoints produced a root mean square error (RMSE) of 5.58 cm in the Z direction. This finding indicates that RTK-equipped UAVs can achieve acceptable error margins even without using GCPs, thereby fulfilling the precision requirements for large-scale mapping.

MFEFNet: A Multi-Scale Feature Information Extraction and Fusion Network for Multi-Scale Object Detection in UAV Aerial Images

Unmanned aerial vehicles (UAVs) are now widely used in many fields. Due to the randomness of UAV flight height and shooting angle, UAV images usually have the following characteristics: many small objects, large changes in object scale, and complex background. Therefore, object detection in UAV aerial images is a very challenging task. To address the challenges posed by these characteristics, this paper proposes a novel UAV image object detection method based on global feature aggregation and context feature extraction named the multi-scale feature information extraction and fusion network (MFEFNet). Specifically, first of all, to extract the feature information of objects more effectively from complex backgrounds, we propose an efficient spatial information extraction (SIEM) module, which combines residual connection to build long-distance feature dependencies and effectively extracts the most useful feature information by building contextual feature relations around objects. Secondly, to improve the feature fusion efficiency and reduce the burden brought by redundant feature fusion networks, we propose a global aggregation progressive feature fusion network (GAFN). This network adopts a three-level adaptive feature fusion method, which can adaptively fuse multi-scale features according to the importance of different feature layers and reduce unnecessary intermediate redundant features by utilizing the adaptive feature fusion module (AFFM). Furthermore, we use the MPDIoU loss function as the bounding-box regression loss function, which not only enhances model robustness to noise but also simplifies the calculation process and improves the final detection efficiency. Finally, the proposed MFEFNet was tested on VisDrone and UAVDT datasets, and the mAP0.5 value increased by 2.7% and 2.2%, respectively.

BDHE-Net: A Novel Building Damage Heterogeneity Enhancement Network for Accurate and Efficient Post-Earthquake Assessment Using Aerial and Remote Sensing Data

Accurate and efficient post-earthquake building damage assessment methods enable key building damage information to be obtained more quickly after an earthquake, providing strong support for rescue and reconstruction efforts. Although many methods have been proposed, most have limited effect on accurately extracting severely damaged and collapsed buildings, and they cannot meet the needs of emergency response and rescue operations. Therefore, in this paper, we develop a novel building damage heterogeneity enhancement network for pixel-level building damage classification of post-earthquake unmanned aerial vehicle (UAV) and remote sensing data. The proposed BDHE-Net includes the following three modules: a data augmentation module (DAM), a building damage attention module (BDAM), and a multilevel feature adaptive fusion module (MFAF), which are used to alleviate the weight deviation of intact and slightly damaged categories during model training, pay attention to the heterogeneous characteristics of damaged buildings, and enhance the extraction of house integrity contour information at different resolutions of the image. In addition, a combined loss function is used to focus more attention on the small number of severely damaged and collapsed classes. The proposed model was tested on remote sensing and UAV images acquired from the Afghanistan and Baoxing earthquakes, and the combined loss function and the role of the three modules were studied. The results show that compared with the state-of-the-art methods, the proposed BDHE-Net achieves the best results, with an F1 score improvement of 6.19–8.22%. By integrating the DBA, BDAM, and MFAF modules and combining the loss functions, the model’s classification accuracy for severely damaged and collapsed categories can be improved.

Weed species classification with UAV imagery and standard CNN models: assessing the frontiers of training and inference phases

Accurate weed species identification is crucial for effective site-specific weed management (SSWM), enabling targeted and timely control measures for each weed in crop field. This study advanced the current approach to species-level weed identification during the early growth stage by integrating unmanned aerial vehicles (UAVs) imagery with standard convolutional neural networks (CNNs) models such as VGG16, Resnet152 and Inception-Resnet-v2. For this, a robust dataset was created with 33,467 labels of weeds (Atriplex patula, Chenopodium album, Convolvulus arvensis, Cyperus rotundus, Lolium rigidum, Portulaca oleracea, Salsola kali, Solanum nigrum) and crops (maize, tomato), which was subjected to different training, validation and test scenarios. Model inputs were adjusted in order to align them with the information represented by the UAV images. Initially, models were developed in balanced scenarios, gradually increasing label numbers to assess their performance. Inception-ResNet-v2 achieved over 90% accuracy with 400 labels, while ResNet152 and VGG16 required 600 and 800 labels, respectively, for similar accuracy. In a more complex and realistic scenarios with unbalanced datasets, Inception-ResNet-v2 outperformed, likely due to its deeper architecture and enhanced capability to capture intricate features and patterns within UAV images. The study emphasized the importance of the minority-to-majority species ratio in unbalanced datasets, which affects minority species classification. To prevent misclassification, it is crucial to determine the right number of labels for CNN model training and validation. Weed maps were generated after species classification using the Faster R-CNN algorithm as an object detector. This advancement in methodology facilitates the precise and efficient implementation of SSWM techniques.

Improving Target Geolocation Accuracy with Multi-View Aerial Images in Long-Range Oblique Photography

Target geolocation in long-range oblique photography (LOROP) is a challenging study due to the fact that measurement errors become more evident with increasing shooting distance, significantly affecting the calculation results. This paper introduces a novel high-accuracy target geolocation method based on multi-view observations. Unlike the usual target geolocation methods, which heavily depend on the accuracy of GNSS (Global Navigation Satellite System) and INS (Inertial Navigation System), the proposed method overcomes these limitations and demonstrates an enhanced effectiveness by utilizing multiple aerial images captured at different locations without any additional supplementary information. In order to achieve this goal, camera optimization is performed to minimize the errors measured by GNSS and INS sensors. We first use feature matching between the images to acquire the matched keypoints, which determines the pixel coordinates of the landmarks in different images. A map-building process is then performed to obtain the spatial positions of these landmarks. With the initial guesses of landmarks, bundle adjustment is used to optimize the camera parameters and the spatial positions of the landmarks. After the camera optimization, a geolocation method based on line-of-sight (LOS) is used to calculate the target geolocation based on the optimized camera parameters. The proposed method is validated through simulation and an experiment utilizing unmanned aerial vehicle (UAV) images, demonstrating its efficiency, robustness, and ability to achieve high-accuracy target geolocation.

Early Drought Detection in Maize Using UAV Images and YOLOv8+

The escalating global climate change significantly impacts the yield and quality of maize, a vital staple crop worldwide, especially during seedling stage droughts. Traditional detection methods are limited by their single-scenario approach, requiring substantial human labor and time, and lack accuracy in the real-time monitoring and precise assessment of drought severity. In this study, a novel early drought detection method for maize based on unmanned aerial vehicle (UAV) images and Yolov8+ is proposed. In the Backbone section, the C2F-Conv module is adopted to reduce model parameters and deployment costs, while incorporating the CA attention mechanism module to effectively capture tiny feature information in the images. The Neck section utilizes the BiFPN fusion architecture and spatial attention mechanism to enhance the model’s ability to recognize small and occluded targets. The Head section introduces an additional 10 × 10 output, integrates loss functions, and enhances accuracy by 1.46%, reduces training time by 30.2%, and improves robustness. The experimental results demonstrate that the improved Yolov8+ model achieves precision and recall rates of approximately 90.6% and 88.7%, respectively. The mAP@50 and mAP@50:95 reach 89.16% and 71.14%, respectively, representing respective increases of 3.9% and 3.3% compared to the original Yolov8. The UAV image detection speed of the model is up to 24.63 ms, with a model size of 13.76 MB, optimized by 31.6% and 28.8% compared to the original model, respectively. In comparison with the Yolov8, Yolov7, and Yolo5s models, the proposed method exhibits varying degrees of superiority in mAP@50, mAP@50:95, and other metrics, utilizing drone imagery and deep learning techniques to truly propel agricultural modernization.

WBGT prediction with high spatial resolution using actual measurement data and data acquired using infrared sensors mounted on UAVs

Spatial understanding of heat stress indices is important to prevent the worsening of heat-related health problems. The objective of this study is to understand heat stress indices with high spatial and temporal resolution using actual measurement data and data acquired using infrared sensors mounted on unmanned aerial vehicles (UAVs). Two study sites were established to examine the relationship between green space and heat stress indices and the versatility of the developed method. The wet bulb globe temperature (WBGT), a heat stress index, was measured on extremely hot days. Thermal infrared, the normalized difference vegetation index, and the normalized difference water index were obtained from UAV images. Extreme gradient boosting, a supervised machine learning method for gradient boosting decision trees, was used to predict the WBGT for the study area using data from actual measurement points as training data. The maximum and average root mean square errors of the predicted WBGT were 2.8 and 1.4, respectively. The second measurement session, which was closest to noon, showed a greater WBGT reduction effect because of vegetation and moisture. This study successfully predicted the WBGT with high spatial and temporal resolution. This achievement is expected to contribute to the prevention of heat stress.

Elimination of Irregular Boundaries and Seams for UAV Image Stitching with a Diffusion Model

Unmanned aerial vehicle (UAV) image stitching refers to the process of combining multiple UAV images into a single large-format, wide-field image, and the stitched image often contains large irregular boundaries and multiple stitching seams. Usually, irregular boundaries are addressed using grid-constrained methods, while seams are optimized through the design of energy functions and penalty terms applied to the pixels at the seams. The above-mentioned two solutions can only address one of the two issues individually and are often limited to pairwise stitching of images. To the best of our knowledge, there is no unified approach that can handle both seams and irregular boundaries in the context of multi-image stitching for UAV images. Considering that addressing irregular boundaries involves completing missing information for regularization and that mitigating seams involves generating images near the stitching seams, both of these challenges can be viewed as instances of a mask-based image completion problem. This paper proposes a UAV image stitching method based on a diffusion model. This method uniformly designs masks for irregular boundaries and stitching seams, and the unconditional score function of the diffusion model is then utilized to reverse the process. Additional manifold gradient constraints are applied to restore masked images, eliminating both irregular boundaries and stitching seams and resulting in higher perceptual quality. The restoration maintains high consistency in texture and semantics. This method not only simultaneously addresses irregular boundaries and stitching seams but also is unaffected by factors such as the number of stitched images, the shape of irregular boundaries, and the distribution of stitching seams, demonstrating its robustness.

Residual learning-based robotic image analysis model for low-voltage distributed photovoltaic fault identification and positioning.

With the fast development of large-scale Photovoltaic (PV) plants, the automatic PV fault identification and positioning have become an important task for the PV intelligent systems, aiming to guarantee the safety, reliability, and productivity of large-scale PV plants. In this paper, we propose a residual learning-based robotic (UAV) image analysis model for low-voltage distributed PV fault identification and positioning. In our target scenario, the unmanned aerial vehicles (UAVs) are deployed to acquire moving images of low-voltage distributed PV power plants. To get desired robustness and accuracy of PV image detection, we integrate residual learning with attention mechanism into the UAV image analysis model based on you only look once v4 (YOLOv4) network. Then, we design the sophisticated multi-scale spatial pyramid fusion and use it to optimize the YOLOv4 network for the nuanced task of fault localization within PV arrays, where the Complete-IOU loss is incorporated in the predictive modeling phase, significantly enhancing the accuracy and efficiency of fault detection. A series of experimental comparisons in terms of the accuracy of fault positioning are conducted, and the experimental results verify the feasibility and effectiveness of the proposed model in dealing with the safety and reliability maintenance of low-voltage distributed PV systems.

Enhancing precision of root-zone soil moisture content prediction in a kiwifruit orchard using UAV multi-spectral image features and ensemble learning

Accurate and real-time monitoring of soil moisture content (SMC) is of utmost importance for effective field irrigation and maximizing crop water productivity. However, a comprehensive investigation into the inversion study for determining suitable combinations of unmanned aerial vehicle (UAV) image features and enhancing the precision of SMC model prediction has yet to be fully validated within a kiwifruit orchard setting. This study addresses this gap by employing a pre-processing method and an optimal band combination algorithm to assess the impact of various combinations of kiwifruit canopy reflectance and fraction vegetation coverage (FVC) features on the sensitivity of root-zone SMC. Furthermore, an optimal ensemble learning (EL) framework was developed to monitor SMC at various root-zone depths (0–10 cm [SMC10], 0–20 cm [SMC20], 0–30 cm [SMC30], 0–40 cm [SMC40], 0–50 cm [SMC50], 0–60 cm [SMC60]). The key findings of this research highlight the successful derivation of 10 wavebands and FVC features, exhibiting a strong correlation with SMC at different root depths. The gradient boosting (GBDT) model demonstrated the exceptional accuracy in estimating SMC10, with an impressive R2 value of 0.963 ± 0.030 and low RMSE values of 0.238 ± 0.111. Similarly, the eXtreme Gradient Boosting (XGBoost) model outperformed in estimating SMC20 to SMC60, with R2 and RMSE values of 0.963 ± 0.024 and 0.117 ± 0.053, respectively. Additionally, the utilization of the optimal EL model allows for digital mapping of SMC at different depths across fruit growth stages, showcasing superior adaptability for SMC30 to SMC60 (with R2 and RMSE of 0.782 ± 0.090 and 0.037 ± 0.011) compared to SMC10 and SMC20 (with R2 and RMSE of 0.765 ± 0.097 and 0.056 ± 0.024). These results underscore the potential of the EL estimation framework in characterizing the spatial distribution of root-zone SMC at the individual kiwifruit plant level.

Monitoring the spatial–temporal distribution of invasive plant in urban water using deep learning and remote sensing technology

Despite water ecosystems being capable of sustaining biodiversity and enhancing the overall resilience of the urban environment, they are highly susceptible to biological invasions. Invasive aquatic plants (IAPs) threaten the natural environment by reducing the diversity of native aquatic plants and animal communities. Detecting IAPs and mapping their distribution is crucial for the protection of urban water ecosystems. This study is the first of its kind to use high-resolution unmanned aerial vehicle (UAV) imagery and deep learning approaches to monitor the expansion of Pistia stratiotes (water lettuce). A DJI Matrice 300 RTK was utilized to capture time-series images with a resolution of 0.018 m on the Guanyin Lake, Cangshan District, Fuzhzou, China, during an outbreak and subsequent rapid growth stage of water lettuce. Three deep learning model architectures and three backbones were combined to detect water lettuce. Model performance and the ability to generalize the model were evaluated to determine the optimal model for water lettuce detection from time-series high-resolution UAV imagery. Results show that the DeepLabv3 + model with ResNet-34 achieved superior performance in detecting water lettuce from time-series imagery, yielding an average accuracy of 90.24 % (85.33 %≤F1_score ≤ 96.54 %) for water lettuce detection on five different dates. For the UAV image acquired on September 26th, the U-Net model with ResNet-18 yielded the highest accuracy (F1_score = 92.46 %), but it was not the optimal model for multi-temporal water lettuce detection on subsequent dates. The distribution of water lettuce can have large variations at different times, with an average change rate of 49.50 % every two days and the highest change rate up to 60.33 %. The study demonstrates that the combination of UAV imagery and a deep learning model can achieve excellent accuracy for water lettuce monitoring and provide a method to map IAPs in dynamic urban water systems over time.

Mapping large-scale pine wilt disease trees with a lightweight deep-learning model and very high-resolution UAV images

ABSTRACT Pine wilt disease (PWD), caused by pine wood nematodes, has brought a great loss in ecology and economy all over the world. In China, the forest health status is also significantly affected by PWD since 1980, especially in coniferous forests and mixed forest regions. The PWD spreads very fast and can cause a healthy pine wood tree to die within a very short time. An effective way to protect other healthy pine wood trees is to discover PWD trees early. Using unmanned aerial vehicle (UAV) images can help people discover the PWD tree quickly and accurately, and a few automatical methods have been developed to monitor PWD trees including the deep learning methods. Because of the robust spatial-temporal transferability, deep learning methods have become the mainstream algorithms to monitor PWD trees. As we know, the training dataset is the most important material to train a deep learning model. However, there is still a lack of a PWD segmentation dataset so far. To fill this gap, in this paper, we have generated the first open-sourced PWD segmentation dataset based on very high-resolution UAV images to help the community conduct PWD monitor research conveniently. This dataset has 994 training samples, and each sample has visible bands with 512 × 512 pixels, and the spatial resolution of this dataset is 0.05 m. In order to train an advanced segmentation model, we have designed a lightweight deep-learning model for mobile devices or edge devices in this manuscript, named MobileSeg. The main feature of MobileSeg is its decoupling, which uses the re-parameterization technology to improve the model performance. Finally, a large-scale real-world scenario experiment with very high-resolution UAV images was utilized to validate the performance of MobileSeg, the experiment result indicated that MobileSeg has achieved the best performance compared with the recent lightweight segmentation models, and the experiment also proved the effectiveness of the proposed PWD dataset.

Relationships of Fire Rate of Spread with Spectral and Geometric Features Derived from UAV-Based Photogrammetric Point Clouds

Unmanned aerial vehicles (UAVs) equipped with RGB, multispectral, or thermal cameras have demonstrated their potential to provide high-resolution data before, during, and after wildfires and prescribed burns. Pre-burn point clouds generated through the photogrammetric processing of UAV images contain geometrical and spectral information of vegetation, while active fire imagery allows for deriving fire behavior metrics. This paper focuses on characterizing the relationship between the fire rate of spread (RoS) in prescribed burns and a set of independent geometrical, spectral, and neighborhood variables extracted from UAV-derived point clouds. For this purpose, different flights were performed before and during the prescribed burning in seven grasslands and open forest plots. Variables extracted from the point cloud were interpolated to a grid, which was sized according to the RoS semivariogram. Random Forest regressions were applied, obtaining up to 0.56 of R2 in the different plots studied. Geometric variables from the point clouds, such as planarity and the spectral normalized blue–red difference index (NBRDI), are related to fire RoS. In analyzing the results, the minimum value of the eigenentropy (Eigenentropy_MIN), the mean value of the planarity (Planarity_MEAN), and percentile 75 of the NBRDI (NBRDI_P75) obtained the highest feature importance. Plot-specific analyses unveiled distinct combinations of geometric and spectral features, although certain features, such as Planarity_MEAN and the mean value of the grid obtained from the standard deviation of the distance between points (Dist_std_MEAN), consistently held high importance across all plots. The relationships between pre-burning UAV data and fire RoS can complement meteorological and topographic variables, enhancing wildfire and prescribed burn models.

Pothole Detection Based on Superpixel Features of Unmanned Aerial Vehicle Images

Pothole Detection Based on Superpixel Features of Unmanned Aerial Vehicle Images

Estimating maize plant height using a crop surface model constructed from UAV RGB images

Plant height (PH) is an essential agronomic trait that can be used to assist in crop breeding pipelines, assess crop productivity and make crop management decisions. Improving the accuracy of the digital terrain model (DTM) and optimising the PH features of the crop surface model obtained from unmanned aerial vehicle (UAV) images contribute to PH estimation. The influence of the fractional vegetation cover (FVC) on DTM reconstruction accuracy was investigated for the first time, and the influence of the view angle (oblique and nadir) and spatial resolution on the accuracy of maize PH estimation was explored. The results show that the accuracy of the DTM constructed using the inverse distance weighted algorithm was significantly influenced by the FVC conditions. Compared with the DTM constructed using UAV images over bare soil, FVC less than 0.4 was necessary for the accurate construction of the DTM, with average estimation errors of 0.15 m in 2018 and 0.09 m in 2019. Compared with the nadir view, the oblique view resulted in a more accurate 3D reconstruction. When the original spatial resolution of 15 mm was upscaled to 20, 30, 60 and 120 mm, a decreasing trend of PH estimation accuracy was observed, with root mean square error increasing from 0.35 to 0.40 m and mean absolute error increasing from 0.30 to 0.36 m. Overall, this study investigated the optimal FVC conditions for accurate DTM construction and the influence of the view angle and spatial resolution on PH estimation based on UAV RGB images.

Rapidly count crop seedling emergence based on waveform Method(WM) using drone imagery at the early stage

Effectively monitoring seedling emergence is critical to identify missing cotton seedling at early stages, allowing the prompt replenishment of the seedlings to maintain crop yield. However, traditional manual inspections have the limitations with low efficiency, poor timeliness, and large counting errors. Although UAV imagery has been applied in seedling monitoring, there are still chance in cotton seedling monitoring under complex environment conditions. Therefore, there is a urgent need for an efficient, rapid, and low-cost method for monitoring early seedling emergence in cotton crops. This study utilizes RGB images of early seedling emergence captured from an unmanned aerial vehicle (UAV) at a flight height of 10 m and a resolution of 0.33 cm in Shihezi, Xinjiang, on 7 and 12 days after seeding (DAS). The modified Hough transform method is constructed for line detection, and crop rows are subsequently extracted from the lines and masked images. After masking, the waveform is extracted from the excess green index (ExG) greyscale map and smoothed to determine the peaks. Seedlings are counted and located by numbering and positioning the peaks, meanwhile their coordinates and distance determine the number and location of missing seedlings. Images with varying resolutions (ranging from 0.33 to 2.5 cm) and intensity changes (0.8 to 1.2) verify the accuracy and efficiency of the proposed waveform method(WM). The results shown that at 7 DAS, the root mean square error (RMSE) is less than 2.0 plants/row, the relative root mean squared error (RRMSE) is less than 2 %, and the coefficient of determination (R2) is stable at 0.95–0.96 when the resolution is below 1.0 cm. The RMSE, RRMSE and R2 fluctuate more when the resolution is less than 1.0 cm. At 12 DAS, the counting accuracy is stable under different conditions, and the RMSE and RRMSE are distributed smoothly over the range of 1.76–4.39 plants/row and 1.29–5.47 % with different resolutions. The R2 is in the range of 0.88–0.97, RRMSE is 1.90–3.39 % and the RMSE is 1.65–2.90 plants/row under different intensities. The results of cotton seedling counting using the waveform method are less affected by image resolution and intensity changes. Therefore, the proposed WM provides efficient, accurate, and automatic counting for cotton seedlings in response to complex weather conditions. Furthermore, it would give a novel approach to accurately countequally spaced row-sown crops based on UAV images.