Kummool, Sart

Preferred name

Kummool, Sart

Main Affiliation

School of Engineering

sart.ku@kmitl.ac.th

Now showing 1 - 6 of 6

Some of the metrics are blocked by your
consent settings
Automatic weighing machine for improving competitiveness of Thai SMEs in seafood cold storage business
(2016-02-10)
Nimsiriwangso, Banyat
;
Kummool, Sart
;
Julsereewong, Amphawan
;
Ukakimaparn, Prapart
Due to sharply rising labor costs in Thailand, many small and medium sized enterprises (SMEs) in various industrial sectors including seafood industry need to improve their competitive capability. The aim of proposed system is to replace workers with automatic machine for upgrading Thai SMEs in seafood cold storage business to survive and maintain their independence. A technique to design and implement an automatic weighing machine for packaging process of fresh whole fish products is presented. The proposed machine is not only specially developed for use in wet environments but also simple to operate and easy to maintain. Its total cost is about one-fifth cheaper than imported machines. Performances of the implemented machine have been observed at a seafood cold storage in Mahachai, Samut Sakhon, Thailand. With automatically weighting operations, the productivity can be increased by more than 1.6 times, while the labor cost can be reduced to 0.625 times, compared to manual operations. In addition, a payback period is approximately equal to 1.14 years.
Some of the metrics are blocked by your
consent settings
Seeded Region Growing based on Localizing Histogram
(2003-10-16)
Sirisantisamrid, Kaset
;
Kummool, Sart
;
Srisuttaboot, Sasi
;
Thepmanee, Teerawat
This article presents segmenting image by seeded region growing and using a technique for selecting the suitable seed points from localizing histogram. Any gray levels on histogram have the maximum frequency (local maximum) in each period, and then defined as seed points of that period. The seed point closely attached with one then combined together as seed region. The obtained seed regions are performed region growing by iterative method until can not be done. Finally, the obtained sub-regions are merging by consideration from the group of average gray level of histogram or arbitrary setting by user. The purposed method of image segmentation compared with other one as shown on the experimental results and its advantage is reduce computational time.
Some of the metrics are blocked by your
consent settings
Improving Accuracy in Image-Recognition Systems with Techniques for Handling Similar-Looking Items in Inventory Management
(2025-01-01)
Smerpitak, Krit
;
Kummool, Sart
;
Asadi, Farzin
;
Ar-Karachaiphong, Chitsanuwit
;
Sopha, Watcharapon
Inventory management is crucial for businesses dealing with many visually similar items. This paper presents the development of an automated board game detection system utilizing computer vision and deep learning to enhance the accuracy and efficiency of tracking visually similar inventory. The dataset used to train the model was collected using a Pi Camera Module 3 mounted on a Raspberry Pi 5, ensuring consistency in captured images. To improve the model's performance, images were annotated using CVAT, and dataset augmentation was applied using Augmentations, incorporating transformations such as flipping, brightness adjustment, blurring, and noise addition. These augmentations allowed the model to learn and differentiate board games under varying environmental conditions accurately. The selected model was YOLOv8m, chosen for its balance between accuracy and computational efficiency. The model was trained for 20 epochs using a learning rate of 0.0005, a batch size of 12, and augmentation techniques such as Mosaic and MixUp. Additionally, overfitting prevention techniques including early stopping, dropout, and weight decay were applied. After training, the model was converted into the ONNX format to optimize it for deployment on embedded devices and then uploaded to the Raspberry Pi 5 for realworld testing. The system was evaluated on 10 visually similar or small-sized board games, with 10 tests per game, totaling 100 trials. The results showed that the model successfully classified all board games without a single error. Moreover, these findings demonstrate that the proposed system effectively identify board games with high accuracy while maintaining real-Time processing capability. The use of data augmentation and model optimization significantly improved object classification performance, making it a promising approach for inventory management in board game cafes and similar environments.
Some of the metrics are blocked by your
consent settings
Simple control valve diagnostic tool for Foundation Fieldbus technology
(2012-01-01)
Kummool, Sart
;
Pongswatd, Sawai
;
Jareonwatthanayothin, Jetsada
This paper presents a simple diagnostic tool for configuration and diagnosis the control valves that are Foundation Fieldbus devices. National Instrument, NI USB-8486 is the hardware which is designed to connect to Foundation Fieldbus segment and Labview software with NI FBUS Communications Manager API is used to develop the diagnostic tool for parameters checking, stroke testing, data logging for signature comparisons and offline analysis in the control valve. The tool not impacts with other Fieldbus devices and controller in the H1 segment. Therefore, the end user can use the tool online and offline in the segment and reduce engineering cost for expensive vendor's asset management software. In additional, the tool have developed to test step response of control valves that can be developed and applied to analyze effective factors or predict the maintenance in higher level. © 2012 SICE.
Some of the metrics are blocked by your
consent settings
Real-time monitoring and reporting alarm system for pH measurement in wet scrubbers
(2016-01-24)
Promsawat, Thipichpon
;
Kummool, Sart
;
Pongswatd, Sawai
;
Julsereewong, Amphawan
In order to improve a maintenance of wet scrubber operation for achieving harmful and polluting gas removal efficiency at a latex glove manufacturer in Thailand, this paper presents a technique to implement the real-time monitoring and reporting alarm system for pH measurement in wet scrubbers. All pH meter outputs are applied to an existing programmable logic controller (PLC) slave located on the master-slave network for production line control. A PLC master is communicated to an operator workstation running the supervisory control and data acquisition (SCADA) software. In addition, measurement data from the SCADA software are automatically transferred to the statistical process control (SPC) software installed in a data historian server. Traditional routine manual data collection and non real-time email notification are replaced by the proposed system, which provides the continuous automatic data collection as well as the real-time display and alarm functions. The automatic and reliable email notification in real time can be also obtained. Experimental results verifying the performance of the proposed system are also included.
Some of the metrics are blocked by your
consent settings
Feedforward hybrid control using foundation fieldbus: A case study of temperature control with DeltaV system
(2018-06-08)
Pannil, Pittaya
;
Jaroenla, Sutee
;
Julsereewong, Amphawan
;
Kummool, Sart
With Foundation Fieldbus (FF) technology, a control strategy can incorporate function blocks running in both host controller and field devices, which is known as hybrid control. This paper focuses on implementation of feedforward hybrid control by utilizing basic function blocks in H1 field devices and advanced function block in host controller to provide a guideline for detailed design to specify of the FF philosophy to be used in a greenfield or brownfield project for maximizing end user achievement. A laboratory-scale FF H1 segment configured and operated on the DeltaV host system for feedforward on temperature control is used as an illustrative case study to analyze three different configurations by assigning control function block into different field devices in terms of the network load for scheduled communications during segment macrocycle. Experimental results obtained from the preferred control loop with minimum scheduled traffic load verify that an improvement in disturbance response can be achieved. Using control block in field device for FF-based hybrid architecture can, therefore, be applied for creating the feedforward control with disturbance anticipation capability.

Kummool, Sart

Filters

Settings

Sort By

Results per page