::國立雲林科技大學電機工程系::

Press Conference

Department of Electrical Engineering Engaged in Artificial Intelligence, with Fruitful Achievements in Deep Learning and Unmanned Aerial Vehicle Image Recognition, Collaborating with Industries and Academia to Solve Practical Problems

A press conference was held at the department on March 20th, 2024, where two faculty members, Professor Wei-Lung Mao and Professor Ching-Ju Chen, were invited to demonstrate their latest research. More than ten media outlets attended the event.

Professor Wei-Lung Mao's lab has developed two novel deep learning technologies, one for tennis logo alignment and the other for resource recycling, with an aim of enhancing productivity with reduced labor. The former utilized automated optical inspection, industrial cameras, and deep learning models to achieve automatic feeding, flipping, positioning, and printing to address issues such as labor fatigue, occupational safety risks, and product defects in the tennis-ball alignment process. The new development provides customers with clear alignment time and detailed information on error during production as well. Given that annual production of tennis balls exceeds 240 million globally, there is a massive market demand, and manufacturers require significant machinery, manpower as well as working hours to meet it. Deployed on actual production lines already, the technology contributed by the principal investigator Mr. Kai-Che Hsu shortens the alignment time for each ball to around six seconds, a marked improvement over traditional methods. As a second presented work during the press conference, Professor Mao's lab has developed a recyclables-sorting system with deep learning and machine vision technology that accurately identifies recyclables of different types on a conveyor belt using a depth camera and YOLO model. Meanwhile a delta robotic arm is driven to transport them to designated bins in real-time. Such application not only saves a considerable amount of labor but also mitigates errors and omissions that may occur in traditional manual sorting.

The application of automation technologies integrated with smart sensors, machine learning, and machine vision plays a crucial role in industrial production processes. Aforementioned studies are to benefit related industries. Deep learning techniques are realized in pragmatic scenarios for improved production efficiency with higher precision yet at lower labor costs, thereby illuminating a means to advance industries towards a more intelligent, efficient future. A short video showcasing part of past work by Professor Mao's lab can be found here.

Apart from foregoing automation advancement, Professor Ching-Ju Chen's lab leverages unmanned aerial vehicles (UAVs) as instruments for data collection, integrating artificial intelligence, information and communication technology, and Internet of Things with sensors to devise expert-level intelligent recognition systems applicable to actual fields. Inspired by the goal of developing UAV-based applications to address field investigations and industry pain points, her lab has been commissioned by government agencies and industries to deal with environmental issues. Industry-academia cooperation projects involve professional training for students, including:

1. Collaboration with the Tainan Animal Epidemic Prevention and Protection Office on AI-based automatic stray dog detection and investigation via UAV aerial imaging: Using UAV imaging techniques to assist personnel in locating stray dogs in the field in real-time, thereby reducing their attacks on pedestrians and minimizing accidents as well as agricultural or fisheries damage.

2. Collaboration with the Biodiversity Research Center of the Ministry of Agriculture on the development of a UAV-based intelligent multisource image recognition system for agricultural landscapes: Developing cutting-edge technology to automatically identify farmland diversity through UAV images to sustain habitats for wildlife. The project won the 2023 Taiwan Biodiversity Award (Silver Award), the 26th Russian Archimedes International Invention Award (Gold Medal and Special Award), and the 18th Seoul International Invention Exhibition Bronze Medal in 2022.

3. Collaboration with the National Institute of Marine Research on the dynamic visualization of urban spatial changes in the former Dayuan Township: Through historical maps, aerial photos, satellite remote sensing images, and UAV high-resolution spatial modeling, the 100-year history of Anping, Tainan, has been vividly brought to life.

4. Collaboration with National Sun Yat-sen University on demonstrative intelligent coastal surveys of basic characteristics of ocean environment: Conducting large-scale coastal environmental investigation with UAVs to achieve comprehensive surveys.

5. Collaboration with Tajen University on the UAV-based intelligent marine environmental quality survey: Instructing UAVs to collect seawater samples for quality testing in coastal areas, reducing the risk to personnel posed in adverse sea conditions during sample collection.

Professor Chen has employed UAVs to carry out remote sensing image processing, spatial information systems, aerial mapping, and 3D modeling, and has been actively engaged in various government commissioned projects on field investigations, orthoimage production, establishment of three-dimensional digital surface model data, aerial remote sensing image recognition and analysis, and agricultural drone applications. In the future, she will continue to focus on the development of aerial remote sensing image recognition technology that lends itself to field investigations, post-disaster field surveys, and work-related records logging. Her led research team keeps aiming at software/hardware co-design and implementation of new methodology on AOI smart image recognition and public safety risk assessment. Meeting UN sustainable development goals, her research caters for government and industrial needs. Videos covering part of Professor Chen's prior research was briefed here.

A group photo of faculty and students

Professor Wei-Lung Mao introducing his tennis logo alignment system

Graduate student Mr. Shu-Han Huang presenting his deep learning-based recyclables-sorting system

Professor Ching-Ju Chen introducing her UAV-based imaging processing applications in various contexts

The Department of Electrical Engineering is dedicated to providing a well-rounded learning and research environment, so as to implement the principle of integrating research achievements into teaching. Through training in independent thinking and hands-on experiences, students develop disciplines to acquire both theoretical knowledge and practical skills that strengthen them to become prospective professionals who can meet the needs of the nation and society on the whole. Faculty and students focus on fields of Power Systems and Power Electronics, Automation and Systems Control, Information and Communications Engineering, and Integrated Circuit and Systems Design. Graduates have promising career in academia and industrial circles. We look forward to comments from all sectors that enrich the programs offered by the department to empower the youth of next-generation.