Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255795
Jupiter Bakakeu, Dominik Kißkalt, J. Franke, S. Baer, H. Klos, J. Peschke
The paper proposes an artificial intelligence-based solution for the efficient operation of a heterogeneous cluster of flexible manufacturing machines with energy generation and storage capabilities in an electricity micro-grid featuring high volatility of electricity prices. The problem of finding the optimal control policy is first formulated as a game-theoretic sequential decision-making problem under uncertainty, where at every time step the uncertainty is characterized by future weather-dependent energy prices, high demand fluctuation, as well as random unexpected disturbances on the factory floor. Because of the parallel interaction of the machines with the grid, the local viewpoints of an agent are non-stationary and non-Markovian. Therefore, traditional methods such as standard reinforcement learning approaches that learn a specialized policy for a single machine are not applicable. To address this problem, we propose a multi-agent actor-critic method that takes into account the policies of other participants to achieve explicit coordination between a large numbers of actors. We show the strength of our approach in mixed cooperative and competitive scenarios where different production machines were able to discover different coordination strategies in order to increase the energy efficiency of the whole factory floor.
{"title":"Multi-Agent Reinforcement Learning for the Energy Optimization of Cyber-Physical Production Systems","authors":"Jupiter Bakakeu, Dominik Kißkalt, J. Franke, S. Baer, H. Klos, J. Peschke","doi":"10.1109/CCECE47787.2020.9255795","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255795","url":null,"abstract":"The paper proposes an artificial intelligence-based solution for the efficient operation of a heterogeneous cluster of flexible manufacturing machines with energy generation and storage capabilities in an electricity micro-grid featuring high volatility of electricity prices. The problem of finding the optimal control policy is first formulated as a game-theoretic sequential decision-making problem under uncertainty, where at every time step the uncertainty is characterized by future weather-dependent energy prices, high demand fluctuation, as well as random unexpected disturbances on the factory floor. Because of the parallel interaction of the machines with the grid, the local viewpoints of an agent are non-stationary and non-Markovian. Therefore, traditional methods such as standard reinforcement learning approaches that learn a specialized policy for a single machine are not applicable. To address this problem, we propose a multi-agent actor-critic method that takes into account the policies of other participants to achieve explicit coordination between a large numbers of actors. We show the strength of our approach in mixed cooperative and competitive scenarios where different production machines were able to discover different coordination strategies in order to increase the energy efficiency of the whole factory floor.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"198 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115492062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255682
Farhana Aklam, W. Osborn
Trip planning queries are considered an integral part of Location Based Services. In this paper, we investigate Sequential Group Trip Planning Queries (SGTPQ). Given a set of starting and destination locations and an ordered sequence of Categories of Interests (COIs) for a group of users, a SGTP query returns the trip route for all users that minimizes the overall travel distance. We propose two approaches: Dynamic Group Trip Planning (DGTP) and Modified Dynamic Group Trip Planning (M-DGTP). The proposed DGTP approach enables users to plan a group trip in a more flexible manner, while the M-DGTP approach optimizes the total travel distance of the group. We compare the results of our proposed strategies with an existing strategy through experimental evaluation.
旅行计划查询被认为是基于位置的服务的一个组成部分。在本文中，我们研究了顺序群旅行计划查询(SGTPQ)。给定一组用户的一组起始和目的地位置以及兴趣类别(Categories of interest, coi)的有序序列，SGTP查询将返回所有用户的行程路线，从而使总行程距离最小化。本文提出了两种方法:动态组团旅行计划(DGTP)和改进动态组团旅行计划(M-DGTP)。建议的DGTP方法使用户能够以更灵活的方式计划团体旅行，而M-DGTP方法优化了团体的总旅行距离。我们通过实验评估比较了我们提出的策略与现有策略的结果。
{"title":"Dynamic Group Trip Planning Queries in Spatial Databases","authors":"Farhana Aklam, W. Osborn","doi":"10.1109/CCECE47787.2020.9255682","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255682","url":null,"abstract":"Trip planning queries are considered an integral part of Location Based Services. In this paper, we investigate Sequential Group Trip Planning Queries (SGTPQ). Given a set of starting and destination locations and an ordered sequence of Categories of Interests (COIs) for a group of users, a SGTP query returns the trip route for all users that minimizes the overall travel distance. We propose two approaches: Dynamic Group Trip Planning (DGTP) and Modified Dynamic Group Trip Planning (M-DGTP). The proposed DGTP approach enables users to plan a group trip in a more flexible manner, while the M-DGTP approach optimizes the total travel distance of the group. We compare the results of our proposed strategies with an existing strategy through experimental evaluation.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"169 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115715532","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255707
H. Ahmadi, M. Armstrong
Electrical equipment installed on high-voltage (HV) transmission structures may require low-voltage (LV) electrical supply from the distribution network. For example, cell sites for communication antennas and warning lights are the most common applications in BC Hydro's system. Bringing the LV supply to the HV structures introduces a number of electrical concerns. The first concern is the transfer of ground potential rise (GPR) from the HV system to the LV system during a ground fault on the transmission structure. The second concern is the induction in the LV system due to the proximity to the HV transmission line. In addition, there could be system impacts that require special attention, such as reduction in circuit-to-circuit separation in multiple-circuit corridors, pole fire on the LV wood poles, etc. This paper discusses technical solutions to mitigate the identified concerns and system impacts. Amongst the possible recommendations, addition of appropriately rated isolation transformers to the LV feeder and improving the electrical grounding on the HV transmission structure are shown to be the most effective methods for preventing the transfer of hazardous potentials to the customers connected to the same LV feeder. The proposed isolation circuit has been tested in a HV laboratory to confirm its effectiveness.
{"title":"Grid-Connected Low-Voltage Power Supply to Equipment on Transmission Line Structures","authors":"H. Ahmadi, M. Armstrong","doi":"10.1109/CCECE47787.2020.9255707","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255707","url":null,"abstract":"Electrical equipment installed on high-voltage (HV) transmission structures may require low-voltage (LV) electrical supply from the distribution network. For example, cell sites for communication antennas and warning lights are the most common applications in BC Hydro's system. Bringing the LV supply to the HV structures introduces a number of electrical concerns. The first concern is the transfer of ground potential rise (GPR) from the HV system to the LV system during a ground fault on the transmission structure. The second concern is the induction in the LV system due to the proximity to the HV transmission line. In addition, there could be system impacts that require special attention, such as reduction in circuit-to-circuit separation in multiple-circuit corridors, pole fire on the LV wood poles, etc. This paper discusses technical solutions to mitigate the identified concerns and system impacts. Amongst the possible recommendations, addition of appropriately rated isolation transformers to the LV feeder and improving the electrical grounding on the HV transmission structure are shown to be the most effective methods for preventing the transfer of hazardous potentials to the customers connected to the same LV feeder. The proposed isolation circuit has been tested in a HV laboratory to confirm its effectiveness.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125020368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255677
Ahmed Alghamdi, W. Chan
Objective measures of speech intelligibility are commonly used as a practical alternative to expensive and laborious listening tests. The extended short-time objective intelligibility (ESTOI) measure has demonstrated high accuracy in predicting the intelligibility of speech corrupted by many types of degradation. In this paper we propose a modified version of ESTOI that is based on the glimpse model of speech perception in noise. Performance assessment against subjective data reveals that the modified ESTOI is equivalent to ESTOI on three data sets and slightly better than ESTOI on two data sets.
{"title":"Modified ESTOI for improving speech intelligibility prediction","authors":"Ahmed Alghamdi, W. Chan","doi":"10.1109/CCECE47787.2020.9255677","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255677","url":null,"abstract":"Objective measures of speech intelligibility are commonly used as a practical alternative to expensive and laborious listening tests. The extended short-time objective intelligibility (ESTOI) measure has demonstrated high accuracy in predicting the intelligibility of speech corrupted by many types of degradation. In this paper we propose a modified version of ESTOI that is based on the glimpse model of speech perception in noise. Performance assessment against subjective data reveals that the modified ESTOI is equivalent to ESTOI on three data sets and slightly better than ESTOI on two data sets.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123638560","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255686
Mohamed Abu Sharkh, Yong Xu, Eric Leyder
Cloud computing is rapidly becoming the standard through which enterprises of all sizes fulfill their computing infrastructure demands. This work aims at exploring the impact that machine learning algorithms can have on Cloud application behavior profiling and prediction. Although classic machine learning algorithms have been used in Cloud Computing context before, cutting-edge algorithms like deep learning (DL) and reinforcement learning (RL) are yet to be convincingly exploited for this specific problem. Despite being a revelation with fields like image processing and speech recognition, these algorithms (deep neural networks for instance) face adoption challenges outside certain topics. There is a high demand for timely research work that dissects these algorithms and develops novel techniques to facilitate seamless adoption for Cloud providers and clients. In this work, we evaluate the efficiency of machine learning algorithms in the Cloud context by applying them to a large scale application resource utilization data set (TU Delft Bitbrains traces). The objective is to design a Cloud application behavior prediction technique based on machine learning predictors. Any improvement on prediction precision has direct impact on key performance indicators for both Cloud providers and Cloud tenants/clients. Experimental results show the potential of our approach to improve Cloud resource scheduling in a Cloud data center.
{"title":"CloudMach: Cloud Computing Application Performance Improvement through Machine Learning","authors":"Mohamed Abu Sharkh, Yong Xu, Eric Leyder","doi":"10.1109/CCECE47787.2020.9255686","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255686","url":null,"abstract":"Cloud computing is rapidly becoming the standard through which enterprises of all sizes fulfill their computing infrastructure demands. This work aims at exploring the impact that machine learning algorithms can have on Cloud application behavior profiling and prediction. Although classic machine learning algorithms have been used in Cloud Computing context before, cutting-edge algorithms like deep learning (DL) and reinforcement learning (RL) are yet to be convincingly exploited for this specific problem. Despite being a revelation with fields like image processing and speech recognition, these algorithms (deep neural networks for instance) face adoption challenges outside certain topics. There is a high demand for timely research work that dissects these algorithms and develops novel techniques to facilitate seamless adoption for Cloud providers and clients. In this work, we evaluate the efficiency of machine learning algorithms in the Cloud context by applying them to a large scale application resource utilization data set (TU Delft Bitbrains traces). The objective is to design a Cloud application behavior prediction technique based on machine learning predictors. Any improvement on prediction precision has direct impact on key performance indicators for both Cloud providers and Cloud tenants/clients. Experimental results show the potential of our approach to improve Cloud resource scheduling in a Cloud data center.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128721716","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255765
J. Mombach, Cristiane B. R. Ferreira, J. P. Félix, R. Salvini, Fabrízzio Soares
Spelling tests for children are a typical activity in primary schools and clinics specialized in child development. Commonly, some child's letters can be mirrored or rotated, and Optical Character Systems (OCRs) do not recognize nonstandard letters. Consequently, automatic evaluation approaches are harmed in this context. Furthermore, depending on the child's age, identifying mirrored or rotated letters can support earlier diagnoses of learning disabilities, such as dyslexia or dysgraphia. Therefore, we propose a method for identifying the occurrence of mirrored and rotated letters in children's spellings. The approach uses image processing techniques to extract letters from paper tests and performs transformations so it can be recognized automatically. Preliminary results indicate a promising approach, reaching an accuracy of 96% for mirrored letters recognition and 98% in rotated letters.
{"title":"Mirrored and Rotated Letters in Children Spellings: An Automatic Analysis Approach","authors":"J. Mombach, Cristiane B. R. Ferreira, J. P. Félix, R. Salvini, Fabrízzio Soares","doi":"10.1109/CCECE47787.2020.9255765","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255765","url":null,"abstract":"Spelling tests for children are a typical activity in primary schools and clinics specialized in child development. Commonly, some child's letters can be mirrored or rotated, and Optical Character Systems (OCRs) do not recognize nonstandard letters. Consequently, automatic evaluation approaches are harmed in this context. Furthermore, depending on the child's age, identifying mirrored or rotated letters can support earlier diagnoses of learning disabilities, such as dyslexia or dysgraphia. Therefore, we propose a method for identifying the occurrence of mirrored and rotated letters in children's spellings. The approach uses image processing techniques to extract letters from paper tests and performs transformations so it can be recognized automatically. Preliminary results indicate a promising approach, reaching an accuracy of 96% for mirrored letters recognition and 98% in rotated letters.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"154 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127513599","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255768
Afonso U. Fonseca, Leandro L. G. Oliveira, J. Mombach, D. Fernandes, R. Salvini, Fabrízzio Soares
Chest radiography is one of the recommended imaging tests by the World Health Organization for childhood pneumonia diagnosis. In computer-aided diagnostic systems where radiography is the main input, its quality is crucial. The presence of foreign artifacts can, therefore, compromise the performance of these systems. In the radiography exam, foreign artifacts are very common, especially in children, due to the ingestion of objects and the need for immobilization of these patients by third parties. Identification tags, shirt buttons, catheters, tubes and in conventional scanned radiographs, fingerprints, tags, noise and inadequate brightness are some of the artifacts present. In this study, we present an efficient and very simple method for detecting and removing artifacts based on common digital image processing operations such as channel subtraction, edge detection, and morphological operations. We describe the proposed method and evaluate its performance in a database of 200 images. We show that it is robust to identify different types of artifacts regardless of their positions on the radiography. A visual inspection was used to measure the errors and the experimental results showed an accuracy of 0.98 and a processing time of about 375ms per image. As a result of this, the method demonstrates to be a very promising pre-processing tool.
{"title":"Foreign Artifacts Detection on Pediatric Chest X-Ray","authors":"Afonso U. Fonseca, Leandro L. G. Oliveira, J. Mombach, D. Fernandes, R. Salvini, Fabrízzio Soares","doi":"10.1109/CCECE47787.2020.9255768","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255768","url":null,"abstract":"Chest radiography is one of the recommended imaging tests by the World Health Organization for childhood pneumonia diagnosis. In computer-aided diagnostic systems where radiography is the main input, its quality is crucial. The presence of foreign artifacts can, therefore, compromise the performance of these systems. In the radiography exam, foreign artifacts are very common, especially in children, due to the ingestion of objects and the need for immobilization of these patients by third parties. Identification tags, shirt buttons, catheters, tubes and in conventional scanned radiographs, fingerprints, tags, noise and inadequate brightness are some of the artifacts present. In this study, we present an efficient and very simple method for detecting and removing artifacts based on common digital image processing operations such as channel subtraction, edge detection, and morphological operations. We describe the proposed method and evaluate its performance in a database of 200 images. We show that it is robust to identify different types of artifacts regardless of their positions on the radiography. A visual inspection was used to measure the errors and the experimental results showed an accuracy of 0.98 and a processing time of about 375ms per image. As a result of this, the method demonstrates to be a very promising pre-processing tool.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"321 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116430146","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255676
A. Biran, A. Jeremic
QRS detection from an electrocardiogram (ECG) is potentially useful tool in many applications such as diagnosing cardiac diseases, bio-identification, bio-encryption, etc. In this paper, we present an automated algorithm for detecting QRS waves and segmenting ECG signal into separate beats using short time Fourier transform (STFT) and multi-channel ECG feature-based classification. We test the performance of our algorithm using ECG signals of 62 subjects from the ECG ID public database. The results show that our method is capable of extracting QRS waves with 99.45% average QRS segmentation accuracy.
{"title":"Automatic QRS Detection and Segmentation Using Short Time Fourier Transform and Feature Fusion","authors":"A. Biran, A. Jeremic","doi":"10.1109/CCECE47787.2020.9255676","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255676","url":null,"abstract":"QRS detection from an electrocardiogram (ECG) is potentially useful tool in many applications such as diagnosing cardiac diseases, bio-identification, bio-encryption, etc. In this paper, we present an automated algorithm for detecting QRS waves and segmenting ECG signal into separate beats using short time Fourier transform (STFT) and multi-channel ECG feature-based classification. We test the performance of our algorithm using ECG signals of 62 subjects from the ECG ID public database. The results show that our method is capable of extracting QRS waves with 99.45% average QRS segmentation accuracy.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126151859","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255773
Mohamed Chetoui, Andy Couturier, M. Akhloufi
Diabetic Retinopathy (DR) is a retinal lesion due to diabetes. Through blood leaks and excess glucose in the blood vessels, pathological lesions including hemorrhages, exudates and microaneurysms (HM, EX, MA) develop in the eye, which may lead to blindness if not timely treated. In this paper, we propose a deep Convolutional Neural Network (CNN) architecture trained to identify Referable Diabetic Retinopathy (RDR) lesions from retinal fundus images. The model uses a pre-trained network with fine-tuned layers, cosine learning rate decay, and warm up. The efficiency of the proposed architecture has been evaluated on eight public datasets. The results show that the proposed architecture obtains state-of-the-art performance using only publicly available datasets. An explainability algorithm was also developed to show the efficiency of the model in detecting RDR signs.
{"title":"Explainable Deep Learning for Referable Diabetic Retinopathy","authors":"Mohamed Chetoui, Andy Couturier, M. Akhloufi","doi":"10.1109/CCECE47787.2020.9255773","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255773","url":null,"abstract":"Diabetic Retinopathy (DR) is a retinal lesion due to diabetes. Through blood leaks and excess glucose in the blood vessels, pathological lesions including hemorrhages, exudates and microaneurysms (HM, EX, MA) develop in the eye, which may lead to blindness if not timely treated. In this paper, we propose a deep Convolutional Neural Network (CNN) architecture trained to identify Referable Diabetic Retinopathy (RDR) lesions from retinal fundus images. The model uses a pre-trained network with fine-tuned layers, cosine learning rate decay, and warm up. The efficiency of the proposed architecture has been evaluated on eight public datasets. The results show that the proposed architecture obtains state-of-the-art performance using only publicly available datasets. An explainability algorithm was also developed to show the efficiency of the model in detecting RDR signs.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126348279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-08-30DOI: 10.1109/CCECE47787.2020.9255827
G. S. Vieira, B. M. Rocha, H. Pedrini, N. M. Sousa, J. C. Lima, R. M. Costa, Fabrízzio Soares
The use of unmanned aerial vehicles (UAV) is expanding rapidly throughout the world. Nowadays, it is common to find some useful applications that use them in both urban and rural environments. Especially in the second case, the UAV is promoting significant changes in traditional agricultural activities. Thus, current technologies have been incorporated into inspection, surveillance, and agricultural management. In this study, we investigated some practical uses of aerial images in rural areas. A new method for allowing a UAV to understand and interpret visual information from static imagery is presented. Tree detection and shadow segmentation are essential requirements for navigation and visual examination purposes. Therefore, our method deals with these tasks to be a starting point to enable a machine to perform visual inspections in production fields. The proposed method uses computer vision techniques such as visual color enhancement, morphological operations, and segmentation approaches. We performed an evaluation of our system based on a dataset with different types of crop areas and pasture lands. Moreover, we assessed our approach to verify tree canopy and shadow detection. We also verified delineating agricultural fields, and segmentation of sunlight exposed vegetation, as well as vegetation areas covered by shadows. Results demonstrate that our approach provides an exciting and robust approach to be adopted in the field.
{"title":"Visual Detection of Productive Crop and Pasture Fields from Aerial Image Analysis","authors":"G. S. Vieira, B. M. Rocha, H. Pedrini, N. M. Sousa, J. C. Lima, R. M. Costa, Fabrízzio Soares","doi":"10.1109/CCECE47787.2020.9255827","DOIUrl":"https://doi.org/10.1109/CCECE47787.2020.9255827","url":null,"abstract":"The use of unmanned aerial vehicles (UAV) is expanding rapidly throughout the world. Nowadays, it is common to find some useful applications that use them in both urban and rural environments. Especially in the second case, the UAV is promoting significant changes in traditional agricultural activities. Thus, current technologies have been incorporated into inspection, surveillance, and agricultural management. In this study, we investigated some practical uses of aerial images in rural areas. A new method for allowing a UAV to understand and interpret visual information from static imagery is presented. Tree detection and shadow segmentation are essential requirements for navigation and visual examination purposes. Therefore, our method deals with these tasks to be a starting point to enable a machine to perform visual inspections in production fields. The proposed method uses computer vision techniques such as visual color enhancement, morphological operations, and segmentation approaches. We performed an evaluation of our system based on a dataset with different types of crop areas and pasture lands. Moreover, we assessed our approach to verify tree canopy and shadow detection. We also verified delineating agricultural fields, and segmentation of sunlight exposed vegetation, as well as vegetation areas covered by shadows. Results demonstrate that our approach provides an exciting and robust approach to be adopted in the field.","PeriodicalId":296506,"journal":{"name":"2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130336738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：