scholarly journals Deep Learning based Improved Automatic Building Extraction from Open-Source High Resolution Unmanned Aerial Vehicle (UAV) Imagery

2021 ◽  
Author(s):  
Chintan Maniyar ◽  
Minakshi Kumar
Keyword(s):  
Deep Learning ◽  
High Resolution ◽  
Open Source ◽  
Unmanned Aerial Vehicle ◽  
Building Extraction ◽  
Aerial Vehicle

scholarly journals A High-Resolution Spatial and Time-Series Labeled Unmanned Aerial Vehicle Image Dataset for Middle-Season Rice

2020 ◽  
Vol 9 (12) ◽  
pp. 728
Author(s):  
Dongbo Zhou ◽  
Shuangjian Liu ◽  
Jie Yu ◽  
Hao Li
Keyword(s):  
Machine Learning ◽  
Time Series ◽  
Deep Learning ◽  
High Resolution ◽  
Unmanned Aerial Vehicle ◽  
Temporal Information ◽  
Learning Networks ◽  
Image Dataset ◽  
Aerial Vehicle ◽  
Uav Image

The existing remote sensing image datasets target the identification of objects, features, or man-made targets but lack the ability to provide the date and spatial information for the same feature in the time-series images. The spatial and temporal information is important for machine learning methods so that networks can be trained to support precision classification, particularly for agricultural applications of specific crops with distinct phenological growth stages. In this paper, we built a high-resolution unmanned aerial vehicle (UAV) image dataset for middle-season rice. We scheduled the UAV data acquisition in five villages of Hubei Province for three years, including 11 or 13 growing stages in each year that were accompanied by the annual agricultural surveying business. We investigated the accuracy of the vector maps for each field block and the precise information regarding the crops in the field by surveying each village and periodically arranging the UAV flight tasks on a weekly basis during the phenological stages. Subsequently, we developed a method to generate the samples automatically. Finally, we built a high-resolution UAV image dataset, including over 500,000 samples with the location and phenological growth stage information, and employed the imagery dataset in several machine learning algorithms for classification. We performed two exams to test our dataset. First, we used four classical deep learning networks for the fine classification of spatial and temporal information. Second, we used typical models to test the land cover on our dataset and compared this with the UCMerced Land Use Dataset and RSSCN7 Dataset. The results showed that the proposed image dataset supported typical deep learning networks in the classification task to identify the location and time of middle-season rice and achieved high accuracy with the public image dataset.

New method for tree stands parameters determination using high resolution images done by unmanned aerial vehicle

2016 ◽  
Author(s):  
А.С. Алексеев ◽  
А.А. Никифоров ◽  
А.А. Михайлова ◽  
М.Р. Вагизов
Keyword(s):  
High Resolution ◽  
Unmanned Aerial Vehicle ◽  
New Method ◽  
Growth And Yield ◽  
Special Treatment ◽  
Resolution Image ◽  
Growing Stock ◽  
High Resolution Image ◽  
Aerial Vehicle ◽  
Area Unit

В связи со старением информационных материалов о состоянии лесов существует потребность в разработке новых методов таксации древостоев, основанных на применении последних научно-технических достижений в области теории структуры и продуктивности древостоев, дистанционных методов изучения лесов, информационных и ГИС технологий. В статье приведены результаты разработки и проверки нового метода определения таксационных характеристик сомкнутых насаждений на основе правила 3/2 и подобных ему правил Хильми и Рейнеке, с одной стороны, и определения числа деревьев на единице площади по снимку сверх высокого разрешения, полученного с помощью БПЛА, с другой. С теоретической точки зрения эта зависимости величин запаса, средней высоты и среднего диаметра от числа стволов на единице площади относятся к классу аллометрических связей, очень часто встречающихся при количественном описании соотношений частей биологических систем разных уровней иерархии, от организмов до экосистем. Параметры аллометрических зависимостей запаса, средних высоты и диаметра от числа стволов на единице площади были определены для основных лесообразующих пород по данным таблиц хода роста нормальных (полных) древостоев с теоретическим показателем степени и затем использованы для расчетов. Число деревьев на единице площади определялось по снимку с разрешением 7,13 см/пиксель, полученному с помощью 4-роторной платформы. Обработка материалов аэрофотосъемки была выполнена в специализированной фотограмметрической системе Agisoft Photoscan. В результате были получены ортофотоплан и цифровая модель поверхности крон деревьев на изучаемую территорию с определением их высот. Для автоматизированной обработки полученных изображений с целью получения значений числа деревьев на единицу площади был создан специализированный скрипт на языке Java. Погрешности определения таксационных характеристик древостоев предлагаемым методом не выше установленных действующими нормативными материалами. Every time there is a demand for new innovative methods of forest resources estimation based on last achievements in theoretical science, remote sensing methods, information and GIS-technologies. In the paper are presented a new method and the results of its application to forest stands growing stock, mean height and diameter determination. The method is based on rule 3/2 and similar Reineke and Hilmy rules, on one hand and high resolution image made by unmanned aerial vehicle, which used for determination of number of trees per area unit, on other. The above rules are well known in quantitative biology as an allometric and widely used for description of different kind of relations in biological systems of various scale: from organisms to ecosystems. Parameters of above allometric relationships between growing stock, mean height and diameter and stems density per area unit was determine on the base of full stock growth and yield tables for main tree species and after used for experimental calculations. The number of trees per area unit was determined after special treatment of high resolution image made by unmanned flying machine. The growing stock, mean height and diameter determined by suggested method was compared with the data of regular forest inventory. Comparison gives positive result and method may be recommended for further development.

scholarly journals Identifying the Branch of Kiwifruit Based on Unmanned Aerial Vehicle (UAV) Images Using Deep Learning Method

Sensors ◽  
2021 ◽  
Vol 21 (13) ◽  
pp. 4442
Author(s):  
Zijie Niu ◽  
Juntao Deng ◽  
Xu Zhang ◽  
Jun Zhang ◽  
Shijia Pan ◽  
...  
Keyword(s):  
Deep Learning ◽  
Unmanned Aerial Vehicle ◽  
Semantic Segmentation ◽  
Dynamic Monitoring ◽  
Support Vector ◽  
Distribution Maps ◽  
Time Operation ◽  
Aerial Vehicle ◽  
Uav Images ◽  
Segmentation Image

It is important to obtain accurate information about kiwifruit vines to monitoring their physiological states and undertake precise orchard operations. However, because vines are small and cling to trellises, and have branches laying on the ground, numerous challenges exist in the acquisition of accurate data for kiwifruit vines. In this paper, a kiwifruit canopy distribution prediction model is proposed on the basis of low-altitude unmanned aerial vehicle (UAV) images and deep learning techniques. First, the location of the kiwifruit plants and vine distribution are extracted from high-precision images collected by UAV. The canopy gradient distribution maps with different noise reduction and distribution effects are generated by modifying the threshold and sampling size using the resampling normalization method. The results showed that the accuracies of the vine segmentation using PSPnet, support vector machine, and random forest classification were 71.2%, 85.8%, and 75.26%, respectively. However, the segmentation image obtained using depth semantic segmentation had a higher signal-to-noise ratio and was closer to the real situation. The average intersection over union of the deep semantic segmentation was more than or equal to 80% in distribution maps, whereas, in traditional machine learning, the average intersection was between 20% and 60%. This indicates the proposed model can quickly extract the vine distribution and plant position, and is thus able to perform dynamic monitoring of orchards to provide real-time operation guidance.

High-resolution monitoring of beach topography and its change using unmanned aerial vehicle imagery

2018 ◽  
Vol 160 ◽  
pp. 103-116 ◽  
Author(s):  
Benqing Chen ◽  
Yanming Yang ◽  
Hongtao Wen ◽  
Hailin Ruan ◽  
Zaiming Zhou ◽  
...  
Keyword(s):  
High Resolution ◽  
Unmanned Aerial Vehicle ◽  
Aerial Vehicle

scholarly journals Building Extraction in Very High Resolution Imagery by Dense-Attention Networks

2018 ◽  
Vol 10 (11) ◽  
pp. 1768 ◽  
Author(s):  
Hui Yang ◽  
Penghai Wu ◽  
Xuedong Yao ◽  
Yanlan Wu ◽  
Biao Wang ◽  
...  
Keyword(s):  
Deep Learning ◽  
High Resolution ◽  
Building Extraction ◽  
Learning Networks ◽  
Feature Maps ◽  
Low Level ◽  
High Resolution Imagery ◽  
Very High Resolution Imagery ◽  
High Level ◽  
Very High

Building extraction from very high resolution (VHR) imagery plays an important role in urban planning, disaster management, navigation, updating geographic databases, and several other geospatial applications. Compared with the traditional building extraction approaches, deep learning networks have recently shown outstanding performance in this task by using both high-level and low-level feature maps. However, it is difficult to utilize different level features rationally with the present deep learning networks. To tackle this problem, a novel network based on DenseNets and the attention mechanism was proposed, called the dense-attention network (DAN). The DAN contains an encoder part and a decoder part which are separately composed of lightweight DenseNets and a spatial attention fusion module. The proposed encoder–decoder architecture can strengthen feature propagation and effectively bring higher-level feature information to suppress the low-level feature and noises. Experimental results based on public international society for photogrammetry and remote sensing (ISPRS) datasets with only red–green–blue (RGB) images demonstrated that the proposed DAN achieved a higher score (96.16% overall accuracy (OA), 92.56% F1 score, 90.56% mean intersection over union (MIOU), less training and response time and higher-quality value) when compared with other deep learning methods.

scholarly journals Wheat Leaf Area Index Prediction Using Data Fusion Based on High-Resolution Unmanned Aerial Vehicle Imagery

2021 ◽  
Author(s):  
Shuang Wu ◽  
Lei Deng ◽  
Lijie Guo ◽  
Yanjie Wu
Keyword(s):  
High Resolution ◽  
Data Fusion ◽  
Leaf Area Index ◽  
Leaf Area ◽  
Unmanned Aerial Vehicle ◽  
Prediction Accuracy ◽  
Sensor Data ◽  
Support Vector ◽  
Area Index ◽  
Aerial Vehicle

Abstract Background: Leaf Area Index (LAI) is half of the amount of leaf area per unit horizontal ground surface area. Consequently, accurate vegetation extraction in remote sensing imagery is critical for LAI estimation. However, most studies do not fully exploit the advantages of Unmanned Aerial Vehicle (UAV) imagery with high spatial resolution, such as not removing the background (soil and shadow, etc.). Furthermore, the advancement of multi-sensor synchronous observation and integration technology allows for the simultaneous collection of canopy spectral, structural, and thermal data, making it possible for data fusion.Methods: To investigate the potential of high-resolution UAV imagery combined with multi-sensor data fusion in LAI estimation. High-resolution UAV imagery was obtained with a multi-sensor integrated MicaSense Altum camera to extract the wheat canopy's spectral, structural, and thermal features. After removing the soil background, all features were fused, and LAI was estimated using Random Forest and Support Vector Machine Regression.Result: The results show that: (1) the soil background reduced the accuracy of the LAI prediction, and soil background could be effectively removed by taking advantage of high-resolution UAV imagery. After removing the soil background, the LAI prediction accuracy improved significantly, R2 raised by about 0.27, and RMSE fell by about 0.476. (2) The fusion of multi-sensor synchronous observation data improved LAI prediction accuracy and achieved the best accuracy (R2 = 0.815 and RMSE = 1.023). (3) When compared to other variables, 23 CHM, NRCT, NDRE, and BLUE are crucial for LAI estimation. Even the simple Multiple Linear Regression model could achieve high prediction accuracy (R2 = 0.679 and RMSE = 1.231), providing inspiration for rapid and efficient LAI prediction.Conclusions: The method of this study can be transferred to other sites with more extensive areas or similar agriculture structures, which will facilitate agricultural production and management.

Sign in / Sign up

Export Citation Format

Share Document