https://ph02.tci-thaijo.org/index.php/isjet/issue/feed INTERNATIONAL SCIENTIFIC JOURNAL OF ENGINEERING AND TECHNOLOGY (ISJET) 2023-11-10T00:00:00+07:00 Assoc. Prof. Dr. Parinya Sanguansat parinyasan@pim.ac.th Open Journal Systems <p>INTERNATIONAL SCIENTIFIC JOURNAL OF ENGINEERING AND TECHNOLOGY (ISJET) publication of the Panyapiwat Institute of Management. has been published on a continuous basis since 2017. It has been certified by the Thai Journal Citation Index Centre (TCI) as being in the Second Group of Journals in Science and Technology.</p> <p>Purpose to publish and disseminate academic articles in the hose in fields of Engineering, Technology, Innovation, Information Technology, Management Information System, Logistics and Transportation, Agricultural Science and Technology, Animal Science and Aquaculture, Food Science, and other areas in Sciences and Technology. Fields for academics, researchers, instructors, students, and the general.</p> <p><strong>Publication Fee:</strong> None</p> <p><strong>Frequency of Publication</strong><strong>:</strong></p> <p>Twice a year</p> <ul> <li class="show">The first issue is January-June.</li> <li class="show">The Second issue July-December</li> </ul> <p>ISSN: 2586-8527</p> https://ph02.tci-thaijo.org/index.php/isjet/article/view/246780 A Low-Cost IIoT-enabled Computer Vision-based System for Classifying Defect Types and Severity Levels in Industry 4.0 2022-09-14T16:43:23+07:00 Warut Pannakkong warut@siit.tu.ac.th Panisara Kanjanarut panpanisarakj@gmail.com <p><span class="fontstyle0">Industry 4.0 technologies such as the Industrial Internet of Things (IIoT) and Artificial Intelligence (AI) can assist in automating defect<br />detection and classification processes which are crucial for quality control in the manufacturing industry. However, there is still a barrier to adopting the technologies in Small and Medium Enterprises (SMEs) because of their limited budget. This<br />paper presents a low-cost defect detection and classifcation system and an interactive real-time dashboard monitoring IIoT data utilizing a singleboard computer and mainly open-source software. In the system, workpieces will be classified into non-defective (OK) and defective (NG) workpieces. Then, the NG workpieces will be further classified into defective types and severity levels. The workpiece used in the case study is a sticker on a 4.4 cm diameter bottle cap. The defect types are Off-Color, Missing Details, and Scratches, then each type is divided further into four severity levels. From evaluation, the system can achieve 96% when classifัying as<br />OK/NG and 88% accuracy in classifying defective types and levels. The system’s reliability is 100%. Based on experts’ opinions, the proposed system is relatively low-cost, reliable, and accurate for practical uses. The proposed system can be implemented locally or globally via a cloud server.</span> </p> 2023-11-10T00:00:00+07:00 Copyright (c) 2023 Panyapiwat Institute of Management https://ph02.tci-thaijo.org/index.php/isjet/article/view/246521 Automatic detection of Fake Crowdfunding Projects 2022-08-11T10:31:57+07:00 Qi Li 6372100118@stu.pim.ac.th Jian Qu jianqu@pim.ac.th <p><span class="fontstyle0">There may be fake information in some crowdfunding projects. However, it is difcult for crowdfunding platforms and investors to fnd fake information in crowdfunding projects. At present, many scholars have studied the methods for identifying fake information, but most of them studied how to distinguish fake information from news articles. Therefore, this research focuses on how to identify fake information that may exist in crowdfunding projects. The detection of fake crowdfunding projects includes functions such as keyword extraction, external knowledge extraction, and classification of real and fake projects. To identify possible fake information in the crowdfunding project, we need to understand more about the crowdfunding project by extracting the keywords of the crowdfunding projects. Therefore, this research compared TF-IDF, CKPE, YAKE, RAKE, TextRank4zh, FastTextRank, HarvestText, and BERT pre-training model methods. We used precision, recall, and F1 scores to measure the effectiveness of the keyword extraction method. Then, we obtained features for judging the authenticity of crowdfunding projects by extracting external knowledge of keywords. Finally, projects were classifed using a classifcation algorithm. The validity of this study for the classification of fake crowdfunding projects achieves 83.77% by the NB method in the dataset. </span></p> 2023-11-10T00:00:00+07:00 Copyright (c) 2023 Panyapiwat Institute of Management https://ph02.tci-thaijo.org/index.php/isjet/article/view/247213 Dense-YOLO: A Lightweight Weed Detection Platform Based on MSRCP 2022-10-11T14:31:09+07:00 MingYuan Wang wangmingyuan@kkumail.com Watis Leelapatra watis@kku.ac.th <p><strong>For real-time weed detection needs and the flexibility of deploying model in embedded devices. We proposed a lightweight object detection platform, named Dense-YOLO which is based on Multi-scale retinex with chromaticity preservation(MSRCP) and&nbsp; YOLOv4 architecture. First, we use MSRCP to preprocess original images to provide a foundation for subsequent feature extraction. Second, Depthwise separable convolution(DSC) is used to reduce parameters, makes it suitable for developing on embedded devices. Third, we used K-means++ to optimize the clustering of anchor size. Fourth, DenseNet-121, PANet and SPP modules together constitute Dense-YOLO. Last, we analyze the effectiveness of focal loss. Compared with YOLOv4, mAP is improved by 7.26%, three-quarters of the parameters are removed and 6.1 higher in FPS. </strong></p> 2023-11-10T00:00:00+07:00 Copyright (c) 2023 Panyapiwat Institute of Management https://ph02.tci-thaijo.org/index.php/isjet/article/view/246551 Defect Reduction in Crispy Coconut Rolls Production Process 2022-05-20T11:15:22+07:00 Ploypailin Phrikthim ploypailinphr@gmail.com Paitoon Siri-O-Ran paitoonsir@pim.ac.th <p><strong> <span class="fontstyle0">The objective of this research is to reduce the defect in the production process of crispy coconut rolls. By using the quality control tool (QC Tool) to find the cause and improve the quality of the production process. This research has used the Check Sheet to find the point outside the control line by using the waste control chart (P-Chart) and the Pareto diagram to distinguish the importance of sequential problems with Pareto’s 80: 20 laws for selecting the most defects in the out of shape. Will focus on this one type of defect and apply this waste reduction in crispy coconut rolls. The problem was analyzed with a fishbone diagram to set up measures to solve the problem. The improvement result was able to reduce waste caused by the crispy coconut rolls production process from the previous loss of 5,028 kg accounting for 8.6%, decreased to 2,949 kg accounting for3.4%, a decrease from June to November 2021 can reduce waste due to amorphous piece of works by 2,079 kg, representing a percentage of waste that can be reduced by 58.65%, representing an annual loss of 415,734 baht per year.</span></strong> </p> 2023-11-10T00:00:00+07:00 Copyright (c) 2023 Panyapiwat Institute of Management https://ph02.tci-thaijo.org/index.php/isjet/article/view/245962 Key Success Factors for Implementing Industry 4.0 of Thailand Manufacturing 2022-01-25T11:39:06+07:00 Varuj Mingmankong m6322041036@g.siit.tu.ac.th Chawalit Jeenanunta chawalit@siit.tu.ac.th Rujira Chaysiri rchaysiri@siit.tu.ac.th Yasushi Ueki yasushi_ueki@ide.go.jp <p><strong> <span class="fontstyle0">Thailand’s manufacturing is a crucial component to drive the economy. However, most manufacturers in Thailand face obstacles and challenges to reach Industry 4.0 in their production process, which will help improve their productivity<br />and efficiency. This paper aims to identify the key factors for Thailand’s manufacturing firms to implement Industry 4.0. The success factors were identified by literature review and the Delphi method. First, we conducted a semi-structured<br />interview with 2 experts and used the Delphi method to analyze the potential factors. The Analytic Hierarchy Process (AHP) is applied to find the ranking of the top 5 factors. There are 13 experts completing the questionnaire, and they are at the<br />managerial level related to Industry 4.0. Analytic Hierarchy Process (AHP) was applied to find the relative weight of success factors to rank the importance of the success factors and give concrete guidance to implement Industry 4.0. The top 5<br />factors are leadership vision, support from top management, knowledge of technology by an employee, aligning Industry 4.0 with organization strategy, and the process of digitalization of theindustry.</span> </strong></p> 2023-11-10T00:00:00+07:00 Copyright (c) 2023 Panyapiwat Institute of Management https://ph02.tci-thaijo.org/index.php/isjet/article/view/246315 Time Reduction in Picking the Product a Case Study of Roof Tile Warehouse 2022-05-24T14:53:06+07:00 Pakdee Jaisue pakdeejai@gmail.com Paitoon Siri-O-Ran paitoonsir@pim.ac.th <p><strong> <span class="fontstyle0">The warehouse has a problem with waiting time for picking products. Also, the case study has a problem with finding goods. The stored area of goods has stored at a random location. Operators work redundantly and make it difficult to find products. This paper aims to improve working time by using simulation software. FlexSim simulation software is used to simulate the models and analyze the performance by input data layout with different conditions. Generating 3D models of scenarios for simulation software. In the first scenario, create the random laying goods location model. The data used to be an input based on actual work. In addition, the other two scenarios model solutions to improve search times faster and more conveniently. The result of the simulation is 10039.01 seconds for 5 orders in a random model. Second, the classification model has 9274.78 seconds of time working and is stored together on the same side. Third, the distribution group model stores goods in separate locations. This model has a working time of 8920.49 seconds.</span> <br /></strong></p> 2023-11-10T00:00:00+07:00 Copyright (c) 2023 Panyapiwat Institute of Management