Reference Guide Store Optimization Reference Guide Created: May 2017 Last updated: November 2017 Rev: Final
Table of contents INTRODUCTION 3 2 AXIS PEOPLE COUNTER AND AXIS 3D PEOPLE COUNTER 3 2.1 Examples 3 3 AXIS OCCUPANCY ESTIMATOR 5 3.1 Examples 5 4 BEST PRACTICES FOR AXIS PEOPLE COUNTING AND OCCUPANCY ESTIMATOR APPLICATIONS 6 5 AXIS QUEUE MONITOR 11 5.1 Examples 11 6 BEST PRACTICES FOR AXIS QUEUE MONITOR 12 7 AXIS DEMOGRAPHICS IDENTIFIER 12 7.1 Examples 13 8 BEST PRACTICES FOR AXIS DEMOGRAPHICS IDENTIFIER 14
Introduction Retailers are constantly looking for ways for improving and optimizing their merchandising and operations to maximize profit. They base their decisions on information retrieved from different systems, where network video in combination with various video analytics can be one of the most efficient tools to provide accurate data about customer activity, buying behavior and conversion rate of customers into buyers. This customer insight can help retailers plan long term for optimal store lay-out and demand as well as for short-term staff schedules, target demand and impact of marketing and promotional activities. Axis has developed a broad solution offering for Store Optimization, based on a combination of network cameras, network audio solutions, video analytics and reporting tools. Axis offers a broad range of applications running embedded on the cameras called the AXIS Store Optimization Suite. The purpose of this reference guide is to provide a best practice guideline to obtain the best results from the AXIS Store Optimization Suite for system integrators. For more detailed information, please refer to the product pages of each of the applications at www.axis.com 2 AXIS People Counter and AXIS 3D People Counter AXIS People Counter runs fully embedded on Axis ceiling-mounted network cameras and automatically counts in real time the number of people passing under the camera and in what direction. There are two types of people counter available: AXIS People Counter and AXIS 3D People Counter. Figure 1: AXIS 3D People Counter Figure 2: AXIS People Counter Figure 2: Overview of AXIS People Counter 2.1 Examples There are many different kinds of critical data required by a retailer to optimize their store performance such as: 1. Number of people visiting a store and insights 2. Sales conversion rate 3. Weather impact on visitor traffic 4. Optimal staffing AXIS Store Reporter can be used to represent these key data obtained from the AXIS People Counter application as below.
Figure 3: Graphical Reporting in AXIS Store Reporter, data from AXIS People Counter
3 AXIS Occupancy Estimator AXIS Occupancy Estimator application provides estimation of building occupancy levels. Enabling retailers to get analytics about a physical locations' occupancy and utilization it is growing in importance towards understanding the flow of people within a building, as well as occupancy and other facility trends. Figure 1: Pictorial representation of AXIS Occupancy Estimator For an entertainment venue, convention centre, high street retail shops, and shopping malls, occupancy is an indicator of the revenue opportunity. For transportation centres and healthcare centres, it is a measure of demand and important for service planning. 3.1 Examples Figure 2: Overview of AXIS Occupancy Estimator It s very important for a retailer to get information on the occupancy levels of their store since visitors can become irritated if there is a congestion in the moving part of the retail space. The key information available from AXIS Occupancy Estimator is as follows: 1. Average visit time 2. Average occupancy 3. Weather impact on occupancy AXIS Store Reporter can be used to represent this key information obtained from the AXIS Occupancy Estimator application as below.
Figure 3: Graphical Reporting in AXIS Store Reporter, data from AXIS Occupancy Estimator 4 Best practices for AXIS People Counting and Occupancy Estimator applications AXIS People Counter and AXIS Occupancy Estimator are applications which run on Axis network cameras. They are accurate and efficient tools to improve store operations and customer service. AXIS People Counter application is a fully embedded software module for Axis IP cameras, intended for shops and other environments where you need to count people. The application is designed for retail scenarios where it counts objects with the characteristics of an adult pedestrian. The exact height limitation of the object depends on camera model, camera lens, and the selected counter sensitivity. Best practices In addition to the instructions in the camera s Installation guide, there are some important steps to follow for the application to behave in the expected way: Parameters AXIS People Counter AXIS 3D People Counter Minimum installable height 2.7 meters 2.4 meters Maximum installable height 10 meters 4 meters Wide entrances ( 2.54 m) Recommended Not recommended Shadows, strong sunlight, glare Not recommended Recommended & strong reflections Minimum counting height of people People below 1.1 meters are not counted People below 1.3 meters are not counted Minimum Illumination 80 LUX Not available Internet Internet connection not required Internet connection required only during installation of the application on the camera. Table 1: Comparison of AXIS People Counter and AXIS 3D People Counter
Ceiling height (cm) Counting zone width (cm) Min: 240 172 260 203 280 234 300 266 320 297 340 310 360 291 380 272 Max: 400 254 Table 2: Height and width considerations for AXIS 3D People Counter Please refer to https://www.axis.com/global/en/tools/camera-selector-for-retail-analytics for AXIS People Counter height and width recommendations for camera installation. The camera should be mounted straight down above the point where people should be counted in a vertical fashion and should be facing 90 straight down. Figure 2: Mount the camera facing straight down at a 90 angle and not tilted As a rule-of-thumb, the camera covers an area as wide as the camera s mounting height. For details about a specific camera model, see the camera selector for retail analytics tool available at https://www.axis.com/global/en/tools/camera-selector-for-retail-analytics After installation, the covered area can be increased depending on the camera s zoom setting. If you need to cover a wide entrance with several counters, you can configure them under the Neighbour counters tab. In order to determine how many counters are needed to cover the entire width of the entrance, use the Camera selector for retail analytics tool: https://www.axis.com/global/en/tools/camera-selector-for-retail-analytics
Figure 2: Place the camera straight over the area where people are walking. Figure 3: Make sure there is enough light.
Figure 4: Avoid having direct sunlight into the counting zone. Figure 5: Make sure that the camera is looking straight down and is not tilted.
Figure 6: Ensure that there are not any moving objects like the escalator steps on the people counter zone Figure 7: Make sure that doors are not opened into the counting zone.
5 AXIS Queue Monitor Queue management is an increasingly important issue for retailers across the globe for good reasons: mismatch of till staffing with queue size, creates a poor image of store operations and can harm potential sales. AXIS Queue Monitor application provides average number of people in queue and information about the length of the queue (low, mid or high). The flexibility for the individual retailer is great as the application allows for queue settings to be adopted according to store policy and desired service level. 5.1 Examples Figure 1: Overview for AXIS Queue Monitor Queuing is one of the most critical and emotional parts of a retail operation. The following could be known and achieved using the AXIS Queue Monitor: 1. Duration of a queue 2. Number of people in a queue AXIS Store Reporter can be used to represent the above key information obtained from the AXIS Queue Monitor application as below. Figure 2: Graphical reporting in AXIS Store Reporter, data from AXIS Queue Monitor Note that the representation of the number of people is only an estimate based on the movement in the interest region (queue area) and not the absolute value at any given point in time.
6 Best practices for AXIS Queue Monitor Figure 1: Mount the camera facing straight down 1. The camera should be installed at a high position. 2. A good overview of the scene must be able to be captured from the high position installation. 3. A slight angle is okay, as long as the object measured is about the same size in every position in the live view. 7 AXIS Demographics Identifier AXIS Demographics Identifier application constantly detects and analyses the faces of visitors, and provides the gender of the visitors and their approximate age range. The application is intended for retail and other environments where you want to estimate the gender and age of people passing by. Figure 1: Overview of AXIS Demographics Identifier
7.1 Examples Demographics analysis is an increasingly important tool for retailers. The key information provided by AXIS Demographics Identifier application is as follows: 1. Number of male and female visitors 2. Age of visitors AXIS Store Reporter can be used to represent this key information obtained from the AXIS Demographics Identifier application as below. Figure 2: Graphical Reporting in AXIS Store Reporter, data from AXIS Demographics Identifier
8 Best practices for AXIS Demographics Identifier Figure 1: Pictorial representation of AXIS Demographics Identifier installation 1. The camera should be installed so that it s directed at the faces of the people walking by. 2. The application tolerates an angle of approximately +/- 15 degrees of yaw, roll or pitch from a frontal face. 3. The camera can be installed from the side of an exit, directed at the people passing by. 4. The camera can also be installed so that it has a view from above, mounted at the exit, directed at the people walking out. 5. In a digital signage scenario, the camera could be installed so that the camera looks straight at the person s face. 6. Aim for a setup where the faces passing by are: As sharply focused as possible Rich in contrast As clear as possible Evenly illuminated where reflections and shadows are avoided as much as possible 7. In scenes with strong backlight, resulting in dark faces with low contrast, consider complementing the setup with illumination sources to achieve better face images.