Image recognition for Smart Tourism system

• The algorithm has been trained to identify a wide variety of tourismrelated objects, such as people on the street or different types of vehicles. This information can be of great use in estimating peak traffic times, high affluence of people, whether COVID-19 safety regulations are met, or even analyzing whether to increase the frequency of public transport in a tourist location
• The system stores the tags of the identified objects and the count of objects of the same type in a database, but does not store the images, allowing the privacy of people and vehicles appearing in the images to be maintained
• The information can be visualized in various types of easily customizable graphs and maps, allowing the extraction of a wider range of information

Input(s):

• Video images captured by CCTV or ip cameras
• Static images

Main feature(s):

• Image processing algorithm for object detection and recognition
• Labelling and classification of objects
• Object counting by type and extraction of indicators

Output(s):

• Tags generated by the image analysis and object counting algorithm according to the use case: -- UC1: number of people, handbags, backpacks and suitcases -- UC2: same objects as in UC1 plus, motorcycles, trucks, bicycles, scooters, buses, dogs and horses. It can also check if security distance between people is kept -- UC3: number of people (not) wearing facemasks -- UC4: number of a variety of vehicles (the same vehicles as in UC2), number of groups of different sizes, and bathers
• Generation of indicators and alerts according to each use case that can allow to predict the number of tourist and people restrictions
• Data visualization panels

• For reliable object detection, objects must be within 20 m (approx.) of the camera
• In the case to measure the security distance, the camera has to be in an angle (front plane) that can measure properly the distance

Local business owners, municipal authorities, urban and city planners, service providers such as transport service providers, hotel business.

To be determined.