How Video Analytics Can Be Used in Winter Destinations

Why are winter destinations challenging? 

The detection of objects is the main part of video analytics, a task that needs performing before objects can be counted or pick times can be analyzed. Winter destinations are challenging because they are different to normal use cases and include different problems.

Usual use cases are where people go on a daily basis, like train stations or supermarkets. These use cases mostly have good infrastructure, which winter destinations don´t always have.

Why can the detection of objects be a problem in winter destinations?

Benefits of usual cases compared to winter destinations are that they are often indoors and thus have better lighting, especially at night when street lamps are the only source of light outside. Darkness is one of the primary burdens for an exact detection. 

Winter destinations highly depend on the weather conditions; rain or snow harms the quality of the camera stream, as does an excess of sun exposure. Wind could blow leaves or other light objects in front of the camera which falsifies the analysis. Also animals, especially birds could change the result of the analysis.

The quantity of people is also an influencing factor, the wider the space is the more persons fit in - which is often the case outside.

Linked to that are the missing opportunities to attach the camera to, because of the mostly wide spaces. That makes the detection or video streams in general harder, because it influences the view and the quality of the detection.

What is a detection?

As already mentioned above, detection of objects is the main part of video analytics. It is the cornerstone for any analysis, because for analytics object counting is necessary and objects have to be detected before they can be counted. The different objects are listed below under the heading “Model Variant”.

When a person or other objects are detected a box appears and covers it, this box exists for anonymization. Each object has its own identification number.

Applications are shown in the overview as a detection, a detection is basically a visualization.

Visualizations

Detection is one of the possible visualizations, Isarsoft Perception offers six different visualizations all of them have different benefits. There are four different heat maps, a trajectory map and of course the detection.

If you are curious about the individual benefits of each visualization, read our Blogpost about Visualizations.

Model Variant

Model Variant is a step that needs to be undertaken, when creating an Application of the type Object Count or Object Flow, which makes the classification of objects possible. For the Crowd Count Application, the only class type would be 'Person'.Therefore, there would be no need to choose.

Within this step you have to choose which object class you want. The list is as follows:

• Person
• Person & Head
• Traffic
• Public Transport

Isarsoft offers tiny versions in comparison to the normal variant, the tiny version requires fewer computing resources and may be less precise, especially over longer distances. It is available for all variants except for public transport.

People counting 

Object Counting builds the base of Isarsoft Perception and People Counting is a big part of it. Isarsoft Perception will then count the detected objects, depending on the selected object class. To get a counting result of a specific area, a zone must be created. If a specific tally needs to be drawn up for 'objects' entering or leaving a area, the software allows the user to create one or multiple lines for the purpose.

We already have some articles about counting on the website whether in trains, for major events or counting with heat maps.  

What is an Application

An Application must be created to use a camera, the Application makes every camera count.

Applications enable objects to get counted and the important KPI´s get filled with precise data. For an Application Lines and Zones can be drawn. Watch our video if you don´t know how to create an Application.

KPIs  

KPIs are Key Performance Indicators, by the application generated metrics. These indicators are the precise data of counting objects. KPIs are the features owners and operators look at, because they get information about the utilisation of the facility, which are worthy insights. Insights into happenings which perhaps would have been invisible for the human eye otherwise.

According to the KPIs, operating decisions are made. These decisions could be budget-related, time-saving or vision-oriented.

How can I create an Application

The process of creating an Application is really easy, but because there are three different Application Types, there are three different instructions. Click on the the type you want to create:

• Object Flow 
• Object Count 
• Crowd Count

Potential use cases for Video Analytics in winter destinations

Accident Prevention

Enhanced safety is crucial in winter destinations, with the assurance of security being one of the primary boosts to tourism.

Isarsoft Perception is trained to detect unusual behavior, it can recognize a situation wherein a person lays or spends an over average long time in one place. This is usually beneficial to detect homeless persons, but in a potential use case in a winter destination this is helpful to detect injured persons. This functionality only works with a preset alarm, with the condition dwell time and a chosen time limit.

Because of the low temperature in most winter destinations, time limits should be shortened to prevent physical damage of the person but also to prevent accidents with other skiers or snowboarders for example.

Analyze Behavior

Video analytics in winter destinations is not only helpful for accident prevention or fast emergency assistance but also to speed up processes. For example Queue Management Systems at Christmas markets, or ski lifts.

Accuracy of Isarsoft Perception

The accuracy of a detection and thus the collected data depends on certain conditions. Winter destinations can be quite challenging, as mentioned previously, due to variables such as snow, poor visibility, excessive sun exposure, etc. To get a better understanding of what influences the detection we listed it:

• Distance
• Light
• Angle of the camera
• Disturbance free recording (no rotation, no things in front of the camera)

More about Isarsoft

With Isarsoft Perception, your camera systems become part of your business intelligence. Whether the goal is to increase efficiency, customer satisfaction or safety, Isarsoft Perception provides the insights needed for informed decisions.

Contact us, to learn more about how to turn security cameras into intelligent sensors.