@article{ author = {SabzaliYameqani, Ali and Alesheikh, Ali Asghar and Majidi, Mostaf}, title = {Development of an Ensemble Learning Approach for Soybean Yield Prediction using Satellite and Meteorological Data}, abstract ={Accurate crop yield estimation is important for many agricultural issues, including agricultural management, national food policies, and international crop trade. For this purpose, various methods are used to predict product performance, and the use of satellite images increases every day. Satellite remote sensing techniques that cover large areas continuously can help in more accurate assessment of crop yields. This research develops an optimal model for predicting soybean yield in the Midwest region of the United States. The ensemble learning hybrid model was tested using satellite images and meteorological data during the dominant growth period. In particular, the Golden Eagle Optimization (GEO) algorithm was used to adjust the hyper-parameters of the XGBoost model to provide the best possible configuration to improve accuracy. The results showed that the GEO-XGBoost model had good results for soybean crop (R equal to 0.9377 and RMSE equal to 0.2394 tons/ha). These results show that the optimized GEO-XGBoost model can provide accurate predictions for soybean yield under different weather conditions and can also be extended to predict other crops in the future.}, Keywords = {Ensemble Learning, Yield Prediction, Soybean, XGBoost, Golden Eagle Optimization}, volume = {14}, Number = {1}, pages = {1-12}, publisher = {انجمن علمي مهندسي نقشه برداري و ژئوماتيک ايران}, doi = {10.61186/jgst.14.1.1}, url = {http://jgst.issgeac.ir/article-1-1193-en.html}, eprint = {http://jgst.issgeac.ir/article-1-1193-en.pdf}, journal = {Journal of Geomatics Science and Technology}, issn = {2322-102X}, eissn = {}, year = {2024} } @article{ author = {Vafaeinejad, Alireza and Soleimanimatin, sahar and Haery, s}, title = {A combined approach to geostatistical models and spatial information systems to evaluate the role of spatial parameters in the prevalence of cancer in Iran: a case study: bladder cancer}, abstract ={According to the statistics of the World Health Organization, about 7.6 million people die from cancer every year, which is 70% of all deaths in low- and middle-income countries. It is also estimated that this number will reach more than 13 million people by 2030. Since the rate of cancer is increasing and the issues related to the spread of diseases, especially cancer, are directly related to geography and environment, this issue has been considered for a long time. The aim of this study is to develop statistical location models and use spatial information system to evaluate the role of spatial parameters and environmental factors in the prevalence of bladder cancer in the provinces of Iran. According to the statistical analysis of bladder cancer data in Iranian provinces, it was found that Yazd and Sistan-Baluchestan provinces have the highest and lowest incidence rates of bladder cancer in Iran, respectively. Also, by setting a minimum value and considering the global average of bladder cancer as a threshold, it was found that the rate of bladder cancer is low in 23% of the provinces, average in 30% of them, and higher than the global average in the remaining 47%. Bladder cancer, as one of the common public health problems in Iran, has a heterogeneous distribution in the country. Using random forest algorithm, this study examines the spatial distribution of bladder cancer in Iran and predicts its incidence in different regions. The results show that the random forest algorithm predicts the incidence of bladder cancer throughout Iran with considerable accuracy. Running the model on about 30% of the test data showed that the actual value of the data was predicted with an accuracy of 0.62. Finally, by implementing a statistical model, the risk factors of this disease are ranked and the rate of bladder cancer in the country is predicted. According to the results of this study, because the hot spots of bladder cancer are located in the central provinces of Iran, especially Isfahan, Chaharmahal and Bakhtiari provinces, it is suggested to conduct more detailed medical and statistical research in this field.}, Keywords = {Spatial Information System, Geostatistics, Bladder Cancer, Iran}, volume = {14}, Number = {1}, pages = {13-24}, publisher = {انجمن علمي مهندسي نقشه برداري و ژئوماتيک ايران}, doi = {10.61186/jgst.14.1.13}, url = {http://jgst.issgeac.ir/article-1-1186-en.html}, eprint = {http://jgst.issgeac.ir/article-1-1186-en.pdf}, journal = {Journal of Geomatics Science and Technology}, issn = {2322-102X}, eissn = {}, year = {2024} } @article{ author = {Bagheri, Fateme and milan, Asghar}, title = {Using deep learning-based classification methods for interpreting brain MRI images for tumor diagnosis}, abstract ={The classification of brain tumors is very important for evaluating and diagnosing the type of tumors and making decisions for treatment according to the stages of disease progression. Many imaging techniques are used to diagnose brain tumors. However, the MRI method is superior compared to other methods due to better image quality and not relying on ionizing radiation. It is obvious that the more accurate the interpretation is, the more it will help the treatment process, and for this purpose, image classification methods that are widely used in remote sensing can be used. Deep learning is a sub-branch of machine learning, and in recent years, it has had a remarkable performance, especially in the topics of image classification and segmentation. In this article, a deep learning model based on a convolutional neural network is proposed to classify different types of brain tumor using a dataset that classifies tumors into meningioma, glioma, and pituitary. MRI imaging methods have different protocols, in this research, the images obtained based on the T1 protocol with a total of 3064 images, which include the images of 233 patients, were used. With the proposed network structure, the overall accuracy of 97.41% was obtained for the data set. The research results show the ability of the model for brain tumor classification purposes.  }, Keywords = {Brain Tumor, Convolutional Neural Network, Data Augmentation, Deep Learning}, volume = {14}, Number = {1}, pages = {25-36}, publisher = {انجمن علمي مهندسي نقشه برداري و ژئوماتيک ايران}, doi = {10.61186/jgst.14.1.25}, url = {http://jgst.issgeac.ir/article-1-1184-en.html}, eprint = {http://jgst.issgeac.ir/article-1-1184-en.pdf}, journal = {Journal of Geomatics Science and Technology}, issn = {2322-102X}, eissn = {}, year = {2024} } @article{ author = {Zarei, Seyyed Mojtaba and TajFirooz, Bahman and Abadpour, Sepideh and Pasandeh, Nader and KouroshNiya, Ali}, title = {Examining Precision of Multibeam and Single Beam Echo-Sounder Data Through Modeling the Differences (Case Study: Bushehr Port)}, abstract ={Hydrographic surveying in ports and seas and obtaining accurate maps of waterway bottoms is of great importance due to its role in safety of navigation and inspection of marine regions. Utilizing multi-beam eco-sounders instead of single-beam ones in bathymetry operations results in full bathymetric coverage, quality improvement as well as decrease in data acquisition time. In multi-beam bathymetry, the depths on the swath edges are affected by the complexity of used sensors. Therefore, modeling and comparing the data obtained by various sensors will result in data quality and accuracy improvement in addition to optimizing the data volume. In order to examine the difference between the data obtained by single-beam eco-sounders and the data obtained by multi-beam eco-sounders, multi-beam data were fed to the mathematical model using average sorting, after applying necessary corrections such as diffraction of sound in water, vessel rotation errors, etc. in the first step. Then in the second step, control point depths which are the data gathered by single-beam sensors were interpolated into the surface models and their values were estimated. In the third step, these estimations were compared using statistical measures. The result demonstrates that the average depth difference between the two sensors is 0.03 m, the standard deviation equals 0.08 m with 98% confidence interval and the root mean square error is 0.21 m. The research reveals that the precision of the performed survey using multi-beam echosounder matches national standards and special order in S-44 standard of International Hydrographic Organization. Furthermore, considering the field operational time, utilizing multi-beam sensors will result in the reduction of construction costs, more bathymetric coverage, and better precisions.}, Keywords = {Single-beam echosounder, Multibeam echosounder, Hydrography standard, International Hydrographic Organization}, volume = {14}, Number = {1}, pages = {37-50}, publisher = {انجمن علمي مهندسي نقشه برداري و ژئوماتيک ايران}, doi = {10.61186/jgst.14.1.37}, url = {http://jgst.issgeac.ir/article-1-1182-en.html}, eprint = {http://jgst.issgeac.ir/article-1-1182-en.pdf}, journal = {Journal of Geomatics Science and Technology}, issn = {2322-102X}, eissn = {}, year = {2024} } @article{ author = {Malek, Mohammad Reza and Khalili, Mohammad and Moradi, Ghob}, title = {Applying Smart Restrictions In The Face of Epidemic Diseases (Case study of COVID-19)}, abstract ={The spread of Covid-19 has become one of the most important challenges in the world. This disease spread rapidly in the world and involved all countries. Most countries tried to deal with it with non-pharmacological interventions, including the imposition of coronavirus restrictions. Predicting this disease is one of the basic ways to deal with it. By predicting the condition of this disease, control measures to deal with this disease can be better managed. Also, in various studies, factors affecting the spread of this disease have been investigated. Paying attention to these factors can play a significant role in improving the management of this disease. Smart application of restrictions for urban management can prevent social, economic, and environmental losses in addition to controlling the spread of disease. Therefore, a two-stage model was designed for the intelligent application of restrictions. In the first stage, the adaptive neural fuzzy system was used to predict the disease, and in the second stage, the fuzzy expert system was used to apply the influence of the influencing factors on the disease outbreak. Finally, the created vulnerability map was used to apply restrictions. For the first step, the results of the comparison of the root mean square error for three clustering methods showed that the combined clustering with a value of 0.53294 works better than the two fuzzy and reduced clustering methods. In the second stage, factors of population density, elderly population density, spatial displacement, distance from busy places, access to medical resources, and cases of this disease were used and a vulnerability map was calculated and produced for the considered period. This map was used to decide the smart application of restrictions for decision-making bodies, and the map of smart application of restrictions was produced for four education and training bodies, the municipality, the health organization, and the governor's office. The results show that despite the application of complete measures for all works on December 30, 2019, in Qom city, some areas of dangerous situation were not areas without the need to apply restrictions in these areas. In education, there was no need to close 58% of the province's schools in this history, and in the municipality, there was no need to close 40% of the parks and gardens, and they could continue their activities by following health protocols.  }, Keywords = {Epidemic, Covid-19-, Smart job restrictions, Adaptive neural fuzzy system -ANFIS, Fuzzy expert system}, volume = {14}, Number = {1}, pages = {51-62}, publisher = {انجمن علمي مهندسي نقشه برداري و ژئوماتيک ايران}, doi = {10.61186/jgst.14.1.51}, url = {http://jgst.issgeac.ir/article-1-1178-en.html}, eprint = {http://jgst.issgeac.ir/article-1-1178-en.pdf}, journal = {Journal of Geomatics Science and Technology}, issn = {2322-102X}, eissn = {}, year = {2024} } @article{ author = {Aghili, Mohammad Ebrahim and Aghabalaei, Amir}, title = {Target Detection in Panchromatic Remotely Sensed Images Using Deep Learning Algorithms: A Review}, abstract ={In recent years, rapid advancements have been made in the use of deep learning for the detection of targets in high-resolution panchromatic remotely sensed images. This study reviews ongoing research in this field and expresses the innovative findings of existing methods. A wide range of desired detections including buildings, vehicles, aircraft, and ships is covered. Unlike traditional methods, which rely on handcrafted features, deep neural networks provide high-level feature learning and target discrimination capabilities. End-to-end training enables networks to directly learn significant representations from images for accurate recognition. Important architectures reviewed include Faster R-CNN, SSD, and YOLO, which are often customized for target detection. In addition to the direct use of panchromatic images, they are used to improve detection performance by integration with other images such as multispectral and infrared. Utilizing spatial information in panchromatic images has increased the accuracy of target detection. Mechanisms of attention, dense connections, and the combination of multi-scale features have been examined to enhance feature learning and improve results in complex scenes. Data augmentation strategies, transfer learning, exploration of low-quality samples, and model adaptability have been employed to manage limited labeled images and network training processes. Overall, advanced deep learning methods have achieved high accuracy in detecting various types of targets. However, challenges in managing small, dense, and ambiguous targets remain. Continuous learning of deep networks for new targets and model adaptation are the directions of recent methods. This review covers many of the deep learning detection methods and provides insights to steer future research towards practical and powerful detectors for target recognition in high-resolution panchromatic images.}, Keywords = {Panchromatic image, Target detection, Deep learning, Spatial features, Spatial resolution enhancement}, volume = {14}, Number = {1}, pages = {63-83}, publisher = {انجمن علمي مهندسي نقشه برداري و ژئوماتيک ايران}, doi = {10.61186/jgst.14.1.63}, url = {http://jgst.issgeac.ir/article-1-1179-en.html}, eprint = {http://jgst.issgeac.ir/article-1-1179-en.pdf}, journal = {Journal of Geomatics Science and Technology}, issn = {2322-102X}, eissn = {}, year = {2024} } @article{ author = {garousi, amirreza and hosseininaveh, ali and latifi, hoom}, title = {Monitoring deforestation in Arsbaran Biosphere Reserve using multi-temporal satellite images based on the refined U-Net network}, abstract ={Deforestation remains a significant concern regarding climate change and biodiversity conservation. At the same time, the development of new image processing techniques and wide access to high spatial and temporal resolution satellite imagery have created unique conditions for monitoring deforestation. This is particularly important in areas such as the Arasbaran Biosphere Reserve. Existing methods for monitoring deforestation rely on a combination of visual inspection, spectral profiles, statistics, and machine learning techniques. Given recent advances in image processing using Convolutional Neural Networks (CNNs), this study aims to evaluate the performance of a refined U-Net architecture for identifying forest cover to monitor deforestation in multi-temporal satellite images. In this regard, a deep learning model for monitoring deforestation in the Arasbaran Biosphere Reserve based on the classification of Landsat satellite images from 2000 to 2022 was developed. In this study, the Normalized Difference Vegetation Index (NDVI) was used to create masks, which were then visually corrected. Additionally, the refined U-Net model presented was compared with Random Forest and Artificial Neural Network (ANN) models. The results showed that the refined U-Net outperformed traditional methods in classifying images into forest or non-forest categories, resulting in an overall accuracy of 96.53%, a kappa coefficient of 91.55%, an F1 score of 94.68%, and an IoU of 90.04%. The proposed model can accurately show forest changes and estimate the amounts of forest area increase or decrease. Overall, it was observed that the forest area in Arasbaran increased during the 2000-2022 period. This research indicates that using a refined U-Net can be an effective tool for sustainable forest resource monitoring and management}, Keywords = {Monitoring, Deforestation, Remote Sensing, Arasbaran, U-Net Convolutional Neural Network}, volume = {14}, Number = {1}, pages = {85-103}, publisher = {انجمن علمي مهندسي نقشه برداري و ژئوماتيک ايران}, doi = {10.61186/jgst.14.1.85}, url = {http://jgst.issgeac.ir/article-1-1189-en.html}, eprint = {http://jgst.issgeac.ir/article-1-1189-en.pdf}, journal = {Journal of Geomatics Science and Technology}, issn = {2322-102X}, eissn = {}, year = {2024} }