After enthusiastic participation in the "Creating Ideas: Question Submission Phase," the judges selected ten thematic field proposals related to Tainan City: five on smart transportation, three on smart tourism, and two on smart epidemic prevention. The detailed topics are shown below.
We welcome everyone to sign up for "Concept Creation: Problem Solving Stage - Book Review," where you can unleash the infinite possibilities of creativity and AI technology by submitting your team's application to solve problems and propose the best possible solutions.
The application period for "Concept Creation: Problem Solving Stage - Book Review" is from July 18th to August 14th. The winning team will receive an award certificate from the city government, a TWCC cloud resource quota of NT$500,000 per team, 2TB of ASUS cloud storage per person per year, a job placement opportunity with ASUS Group, and a collaboration opportunity with the Tainan City Government on AI smart applications. This is a rare opportunity, so please seize it!
Smart Transportation
1.
- Proposal Title: Smart Pedestrian Flow Hotspot Map
- Proposal Description: Whenever the number of tourists increases during holidays, traffic jams may occur at some popular locations due to the large crowds. If we can use IoT or telecommunications data to collect statistics on the current and even predict the future crowd gathering situation, people who want to go there can go to other less crowded places first, so that the crowds do not become too concentrated and cause traffic congestion.
2.
- Proposal Title: AI-Powered Intelligent Transportation Management System Based on Drone Aerial Imagery
- Proposal Description: In metropolitan areas with complex traffic environments, roadside detectors using technologies such as video, microwave radar, infrared, and loop detectors are commonly used to obtain real-time traffic information. The Transportation Bureau began piloting drones to collect traffic image data at the end of 2021, using AI recognition software to calculate traffic data such as turning distances and stop/wait line lengths. This data is expected to be used for traffic management monitoring and decision-making in the future. However, because the accuracy of the recognition will affect the subsequent application of intelligent transportation systems, the following issues still need to be addressed:
(1) The trade-off between flight altitude and visibility.
(2) High-altitude teleportation of drones.
(3) Specification recommendations for unmanned aerial vehicles (including cameras).
(4) Regarding the applicability and immediacy of combining drones with AI image recognition in traffic management, are there more practical application methods?
3.
- Proposal Title: Image-Based Identification in Road Excavation Cases
- Proposal Description: For road excavation cases, both at the time of application and at the time of closure, the applying unit provides relevant image data for reference or evidence in the execution of the case. However, because there are tens of thousands of road excavation cases every year, and each case involves a considerable number of images that need to be compared or confirmed, the variety and quantity of these cases are substantial, requiring significant manpower and time to process. For example:
(1) When applying for a case, the images of the certification personnel provided by the applicant must be verified.
(2) When the case is closed, such as in the photos of the construction process, the placement and scale of the measuring tools, or the identification of numbers in the images of each stage of construction.
4.
- Proposal Title: AI Assistance for Righteous Circumstances
- Proposal Description: When driving through certain sections of road, one often experiences the unpleasant sensation of being lifted into the air and then crashing back down. On the road, whether it's faded zebra crossings, illegible road markings, damaged crash barrier debris, or protruding roadside trees, all contribute to a sense of insecurity. The proposal aims to use drones and AI technology to analyze uneven road surfaces (potholes, protrusions, manhole covers), illegible road markings due to paint loss, and detect foreign objects around the lanes (protruding tree branches, fallen objects, damaged crash barrier debris), etc., and to analyze the severity of the hazards, promptly notifying relevant authorities for maintenance.
5.
- Proposal Title: Main Road Signal Control System Based on Image Recognition
- Proposal Description: Due to the narrow road widths in Tainan City, mostly two-lane roads in both directions, congestion frequently occurs during rush hour. Therefore, installing detectors at intersections of major arterial roads and collecting ATSPM (Automated Traffic Signal Performance Measurement) data could allow for signal adjustments based on the number of vehicles queuing at intersections, thus alleviating the congestion problem. ATSPM, included in the Every Day Counts 4 technical plan, is defined as a set of performance measurement, data collection, and data analysis tools to support goal- and performance-based traffic signal operation, maintenance, and management aimed at improving the safety, mobility, and efficiency of signalized intersections for all users.
Smart Tourism
6.
- Proposal Title: Scenic Spot Identification
- Proposal Description: To facilitate quick access to images within the bureau's image database, a significant amount of time is currently spent tagging each image during upload. For example, a photo of the golden shower tree at Hutoupi must be tagged with "Xinhua District," "Hutoupi Scenic Area," "Golden Shower Tree," and "Summer Flower Viewing." Using image recognition and similar word analysis technologies would simplify and streamline the upload process. Once the database is more comprehensive, it would also allow the public to search for attractions using images, and the search results could be used as a basis for tourism data analysis. Initially, it's advisable to use a specific "plant" or a unique "landmark" or "building" at the attraction as identification criteria.
7.
- Proposal Title: Tainan City Tourist Guide App
- Proposal Description: Whenever tourists visit Tainan, they can only rely on Google Maps to find attractions or shops. The recommended places or shops are always similar, which eventually leads to a loss of novelty. The proposal hopes to combine AI technology with the attractions that Tainan City wants to promote each year, as well as the top 100 shops and special snacks selected, to make different recommendations for different customer groups. This would not only allow tourists to see a different side of Tainan, but also play a role in attracting more visitors and alleviating traffic congestion on roads that are prone to overcrowding.
8.
- Proposal Title: Analysis of Visitor Groups to Historical Sites
- Proposal Description: The historical sites in Tainan City are important cultural assets and industrial development projects, and are loved by domestic and foreign tourists, making them must-visit tourist attractions. For the operational needs of these historical sites, analyzing the visitor demographics is crucial. Understanding the composition of visitors is beneficial for developing operational and marketing strategies, and for enhancing and optimizing the visitor experience.
Smart epidemic prevention
9.
- Proposal Title: AI-Powered Inspection and Cleaning System
- Proposal Description: The Tainan City Environmental Protection Bureau uses drones for dengue fever patrols in Annan District, which is not only effective but also significantly reduces the time and manpower required. If AI technology could be used to automatically identify stagnant water containers, piles of garbage, and uncovered water containers in aerial images, it would further reduce manpower and allow multiple drones to be airborne simultaneously, with one person monitoring and transmitting data in real time. AI technology could also be used to analyze past dengue ovitrap monitoring information, dengue mosquito density surveys, rainfall, temperature, and other open data to estimate when drone patrols should be deployed.
10.
- Proposal Title: AI Speaking Skills - Screening with Warranty
- Proposal Description: To improve the screening rate of four types of cancer and protect the health of citizens, the Health Bureau has big data data from many years of implementation, which can be used by the team to analyze screening invitations and public response information. It is hoped that AI technology can be used to generate screening invitation scripts with the highest success rate and apply them to SMS, notification letters or telephone interviews to customize the most effective invitation strategies for different groups of people.