scholarly journals Urban Vegetation Mapping from Aerial Imagery Using Explainable AI (XAI)

Sensors ◽  
2021 ◽  
Vol 21 (14) ◽  
pp. 4738
Author(s):  
Abolfazl Abdollahi ◽  
Biswajeet Pradhan
Keyword(s):  
Spectral Characteristics ◽  
Texture Features ◽  
Vegetation Mapping ◽  
Aerial Imagery ◽  
Training Dataset ◽  
Urban Vegetation ◽  
Island Effect ◽  
Remote Sensing Image Classification ◽  
Explainable Ai ◽  
Vegetation Maps

Urban vegetation mapping is critical in many applications, i.e., preserving biodiversity, maintaining ecological balance, and minimizing the urban heat island effect. It is still challenging to extract accurate vegetation covers from aerial imagery using traditional classification approaches, because urban vegetation categories have complex spatial structures and similar spectral properties. Deep neural networks (DNNs) have shown a significant improvement in remote sensing image classification outcomes during the last few years. These methods are promising in this domain, yet unreliable for various reasons, such as the use of irrelevant descriptor features in the building of the models and lack of quality in the labeled image. Explainable AI (XAI) can help us gain insight into these limits and, as a result, adjust the training dataset and model as needed. Thus, in this work, we explain how an explanation model called Shapley additive explanations (SHAP) can be utilized for interpreting the output of the DNN model that is designed for classifying vegetation covers. We want to not only produce high-quality vegetation maps, but also rank the input parameters and select appropriate features for classification. Therefore, we test our method on vegetation mapping from aerial imagery based on spectral and textural features. Texture features can help overcome the limitations of poor spectral resolution in aerial imagery for vegetation mapping. The model was capable of obtaining an overall accuracy (OA) of 94.44% for vegetation cover mapping. The conclusions derived from SHAP plots demonstrate the high contribution of features, such as Hue, Brightness, GLCM_Dissimilarity, GLCM_Homogeneity, and GLCM_Mean to the output of the proposed model for vegetation mapping. Therefore, the study indicates that existing vegetation mapping strategies based only on spectral characteristics are insufficient to appropriately classify vegetation covers.

scholarly journals Can computed tomography-based radiomics potentially discriminate between anterior mediastinal cysts and type B1 and B2 thymomas?

2020 ◽  
Vol 19 (1) ◽  
Author(s):  
Lulu Liu ◽  
Fangxiao Lu ◽  
Peipei Pang ◽  
Guoliang Shao
Keyword(s):  
Computed Tomography ◽  
Sensitivity And Specificity ◽  
Texture Features ◽  
Training Dataset ◽  
Features Selection ◽  
Clinical Factors ◽  
Combined Model ◽  
Test Dataset ◽  
Enhanced Ct ◽  
Mediastinal Cysts

Abstract Background Anterior mediastinal cysts (AMC) are often misdiagnosed as thymomas and undergo surgical resection, which caused unnecessary treatment and medical resource waste. The purpose of this study is to explore potential possibility of computed tomography (CT)-based radiomics for the diagnosis of AMC and type B1 and B2 thymomas. Methods A group of 188 patients with pathologically confirmed AMC (106 cases misdiagnosed as thymomas in CT) and thymomas (82 cases) and underwent routine chest CT from January 2010 to December 2018 were retrospectively analyzed. The lesions were manually delineated using ITK-SNAP software, and radiomics features were performed using the artificial intelligence kit (AK) software. A total of 180 tumour texture features were extracted from enhanced CT and unenhanced CT, respectively. The general test, correlation analysis, and LASSO were used to features selection and then the radiomics signature (radscore) was obtained. The combined model including radscore and independent clinical factors was developed. The model performances were evaluated on discrimination, calibration curve. Results Two radscore models were constructed from the unenhanced and enhanced phases based on the selected four and three features, respectively. The AUC, sensitivity, and specificity of the enhanced radscore model were 0.928, 89.3%, and 83.8% in the training dataset and 0.899, 84.6%, and 87.5% in the test dataset (higher than the unenhanced radscore model). The combined model of enhanced CT including radiomics features and independent clinical factors yielded an AUC, sensitivity and specificity of 0.941, 82.1%, and 94.6% in the training dataset and 0.938, 92.3%, and 87.5% in the test dataset (higher than the unenhanced combined model and enhanced radscore model). Conclusions The study suggested the possibility that the combined model in enhanced CT provided a potential tool to facilitate the differential diagnosis of AMC and type B1 and B2 thymomas.

A primer on mapping vegetation using remote sensing

2001 ◽  
Vol 10 (4) ◽  
pp. 277 ◽  
Author(s):  
Tom Bobbe ◽  
Henry Lachowski ◽  
Paul Maus ◽  
Jerry Greer ◽  
Chuck Dull
Keyword(s):  
Remote Sensing ◽  
Remotely Sensed ◽  
Vegetation Mapping ◽  
Project Planning ◽  
Gis Mapping ◽  
Remotely Sensed Data ◽  
Geospatial Information ◽  
Ecological Principles ◽  
Vegetation Maps ◽  
Key Steps

This paper was presented at the conference ‘Integrating spatial technologies and ecological principles for a new age in fire management’, Boise, Idaho, USA, June 1999 The use of information based upon remotely sensed data is a central factor in our 21st Century society. Scientists in land management agencies especially require accurate and current geospatial information to effectively implement ecosystem management. The increasing need to collect data across diverse landscapes, scales, and ownerships has resulted in a wider application of remote sensing, Geographic Information Systems (GIS) and associated geospatial technologies for natural resource applications. This paper summarizes the use of digital remotely sensed data for vegetation mapping. Key steps in preparing vegetation maps are described. These steps include defining project requirements and classification schemes, use of reference data, classification procedures, and assessing accuracy. The role of field personnel and inventory data is described. Case studies and applications of vegetation mapping on national forest land are also included. remote sensing, GIS, mapping, geospatial, project planning.

scholarly journals STUDY ON REMOTE SENSING IMAGE CHARACTERISTICS OF ECOLOGICAL LAND: CASE STUDY OF ORIGINAL ECOLOGICAL LAND IN THE YELLOW RIVER DELTA

2018 ◽  
Vol XLII-3 ◽  
pp. 35-38
Author(s):  
G. Q. An
Keyword(s):  
Remote Sensing ◽  
Land Use ◽  
Vegetation Index ◽  
Yellow River ◽  
Spectral Characteristics ◽  
Texture Features ◽  
Yellow River Delta ◽  
River Delta ◽  
The Yellow River ◽  
The Yellow River Delta

Takes the Yellow River Delta as an example, this paper studies the characteristics of remote sensing imagery with dominant ecological functional land use types, compares the advantages and disadvantages of different image in interpreting ecological land use, and uses research results to analyse the changing trend of ecological land in the study area in the past 30 years. The main methods include multi-period, different sensor images and different seasonal spectral curves, vegetation index, GIS and data analysis methods. The results show that the main ecological land in the Yellow River Delta included coastal beaches, saline-alkaline lands, and water bodies. These lands have relatively distinct spectral and texture features. The spectral features along the beach show characteristics of absorption in the green band and reflection in the red band. This feature is less affected by the acquisition year, season, and sensor type. Saline-alkali land due to the influence of some saline-alkaline-tolerant plants such as alkali tent, Tamarix and other vegetation, the spectral characteristics have a certain seasonal changes, winter and spring NDVI index is less than the summer and autumn vegetation index. The spectral characteristics of a water body generally decrease rapidly with increasing wavelength, and the reflectance in the red band increases with increasing sediment concentration. In conclusion, according to the spectral characteristics and image texture features of the ecological land in the Yellow River Delta, the accuracy of image interpretation of such ecological land can be improved.

scholarly journals Landslide mapping with multi-scale object-based image analysis – a case study in the Baichi watershed, Taiwan

2011 ◽  
Vol 11 (10) ◽  
pp. 2715-2726 ◽  
Author(s):  
T. Lahousse ◽  
K. T. Chang ◽  
Y. H. Lin
Keyword(s):  
Image Analysis ◽  
Spectral Characteristics ◽  
Satellite Image ◽  
Shallow Landslides ◽  
Training Dataset ◽  
Validation Dataset ◽  
Spectral Difference ◽  
Multi Scale ◽  
Object Based Image Analysis ◽  
Object Based

Abstract. We developed a multi-scale OBIA (object-based image analysis) landslide detection technique to map shallow landslides in the Baichi watershed, Taiwan, after the 2004 Typhoon Aere event. Our semi-automated detection method selected multiple scales through landslide size statistics analysis for successive classification rounds. The detection performance achieved a modified success rate (MSR) of 86.5% with the training dataset and 86% with the validation dataset. This performance level was due to the multi-scale aspect of our methodology, as the MSR for single scale classification was substantially lower, even after spectral difference segmentation, with a maximum of 74%. Our multi-scale technique was capable of detecting landslides of varying sizes, including very small landslides, up to 95 m2. The method presented certain limitations: the thresholds we established for classification were specific to the study area, to the landslide type in the study area, and to the spectral characteristics of the satellite image. Because updating site-specific and image-specific classification thresholds is easy with OBIA software, our multi-scale technique is expected to be useful for mapping shallow landslides at watershed level.

Evaluation of bispectral LIDAR data for urban vegetation mapping

2016 ◽  
Author(s):  
Jean Nabucet ◽  
Laurence Hubert-Moy ◽  
Thomas Corpetti ◽  
Patrick Launeau ◽  
Dimitri Lague ◽  
...  
Keyword(s):  
Vegetation Mapping ◽  
Lidar Data ◽  
Urban Vegetation

scholarly journals A Vegetation Mapping Strategy for Conifer Forests by Combining Airborne LiDAR Data and Aerial Imagery

2015 ◽  
Vol 42 (1) ◽  
pp. 1-15 ◽  
Author(s):  
Yanjun Su ◽  
Qinghua Guo ◽  
Danny L. Fry ◽  
Brandon M. Collins ◽  
Maggi Kelly ◽  
...  
Keyword(s):  
Vegetation Mapping ◽  
Airborne Lidar ◽  
Aerial Imagery ◽  
Lidar Data ◽  
Conifer Forests ◽  
Mapping Strategy ◽  
Airborne Lidar Data

scholarly journals Can Computed Tomography-Based Radiomics Potentially Discrimination Between Anterior Mediastinal Cysts and type B1 and B2 thymomas

2020 ◽  
Author(s):  
Lulu Liu ◽  
Fangxiao Lu ◽  
Peipei Pang ◽  
Guoliang Shao
Keyword(s):  
Computed Tomography ◽  
Sensitivity And Specificity ◽  
Texture Features ◽  
Training Dataset ◽  
Features Selection ◽  
Clinical Factors ◽  
Combined Model ◽  
Test Dataset ◽  
Enhanced Ct ◽  
Mediastinal Cysts

Abstract Background Anterior mediastinal cysts (AMC) are often misdiagnosed as thymomas and undergo surgical resection, which caused unnecessary treatment and medical resources waste. The purpose of this study was to explore potential possibility of computed tomography (CT)-based radiomics for the diagnosis of AMC and type B1 and B2 thymomas. Methods A group of 188 patients with pathologically confirmed AMC (106 cases mischarged as thymomas in CT) and thymomas (82 cases) and underwent routine chest CT from January 2010 to December 2018 were retrospectively analyzed. The lesions were manually delineated using ITK-SNAP software, and radiomics features were performed using the Artificial Intelligence Kit (AK) software. A total of 396 tumor texture features were extracted from enhanced CT and unenhanced CT, respectively. The general test, correlation analysis and LASSO were used to features selection and then the radiomics signature (radscore) were obtained. The combined model including radscore and independent clinical factors were developed. The model performances were evaluated on discrimination, calibration curve. Results Two radscore model were constructed from the unenhanced and enhanced phases based on the selected 4 and 3 features, respectively. The AUC, sensitivity, and specificity of the enhanced radscore model were 0.928, 89.3% and 83.8% in the training dataset and 0.899, 84.6%, 87.5% in the test dataset (higher than the unenhanced radscore model). The combined model of enhanced CT including radiomics features and independent clinical factors yielded an AUC, sensitivity and specificity of 0.941, 82.1%, and 94.6% in the training dataset and 0.938, 92.3%, and 87.5% in the test dataset (higher than the unenhanced combined model and enhanced radscore model). Conclusions The study suggested the possibility that the combined model in enhanced CT provided a potential tool to facilitate the differential diagnosis of AMC and type B1 and B2 thymomas.

Near-Field High-Resolution Maps of the Ridgecrest Earthquakes from Aerial Imagery

2021 ◽  
Author(s):  
Alba M. Rodriguez Padilla ◽  
Mercedes A. Quintana ◽  
Ruth M. Prado ◽  
Brian J. Aguilar ◽  
Thomas A. Shea ◽  
...  
Keyword(s):  
High Resolution ◽  
Hazard Analysis ◽  
Near Field ◽  
Lateral Spreading ◽  
Airborne Lidar ◽  
Aerial Imagery ◽  
Training Dataset ◽  
Surface Rupture ◽  
High Level ◽  
Airborne Lidar Data

Abstract High-resolution maps of surface rupturing earthquakes are essential tools for quantifying rupture hazard, understanding the mechanics of rupture propagation, and interpreting evidence of past earthquakes in the landscape. We present highly detailed maps of five portions of the surface rupture of the 2019 Ridgecrest earthquakes, derived from 5 cm per pixel aerial imagery and 2–20 cm per pixel unmanned aerial vehicle imagery. Our high-resolution maps cover areas of complexity and distributed deformation, sections in which strain is very localized, and areas where the rupture breaks through sediment and bedrock, ensuring sampling of the diverse rupture styles of this earthquake sequence. These maps reveal the near-field deformation of the surface rupture with a high level of detail, resolving the extent of secondary fracturing, lateral spreading, and liquefaction features that are below the resolution of airborne lidar data, field mapping, and geodesy. These data may serve as a machine learning training dataset, and offer opportunities for detailed kinematic analysis and high-resolution probabilistic displacement hazard analysis.

Sign in / Sign up

Export Citation Format

Share Document