Intelligent Person Analysis

Principles

Figure 4-262 Principles

Configuration Process

• The intelligent person analysis function implemented using intelligent cameras can be configured on the LDU or in the camera web system.

  Figure 4-263 shows the configuration process on the LDU. For details about the configuration in the camera web system, see the product documentation of the corresponding camera.

  Figure 4-263 Configuration process
  
• The intelligent person analysis implemented using the can be configured only on the LDU. Figure 4-264 shows the configuration process.
  Figure 4-264 Configuration process
  

Function Description

Table 4-237 describes the camera intelligence requirements.

Application Limitations

• For details about the data retention period and clearance rules, see Data Clearance Rules.
• Intelligent person analysis is not supported for video encoded in the MJPEG format.
• To configure intelligent analysis for cameras on the iClient S100, the camera version must be 9.0.0 or later.

Definition

Personal attributes in video can be detected in real time.

Personal attributes include age, gender, tops color, tops style, physique, glasses, and mask.

Customer Benefits

The system automatically recognizes personal attributes, reducing labor costs, shortening service processing time, and enhancing service efficiency.

Application Scenario

This function is mainly used to configure person and vehicle recognition tasks in key areas such as campus roads and fire passages for personal attribute recognition.

Requirements

Application Limitations

This function is used only for PoC outside the Chinese mainland.

Scenario Description

Table 4-240 describes the support for this feature in different client scenarios.

Scenario Description

• Table 4-241 describes the support for this feature in different client scenarios.

• After local intelligence is enabled or an analysis task is created for a camera, do not modify the attributes of the camera. Otherwise, target recognition and video structuring tasks will become invalid, and you will need to delete and recreate the tasks.

(Optional) Configuring a Person Recognition Task

1. On the iClient S100 home page, choose Maintenance Management > Video Device > Device List.
2. Right-click an HWT-IVS1800 and choose Task Management.
3. Click New Task.
4. Set video structuring global parameters, as shown in Figure 4-268.
   Figure 4-268 Setting video structuring global parameters
   

   Table 4-242 describes the parameters.

   Table 4-242 Parameter description

   Parameter	Description
   Camera	Select the camera for which you want to create an analysis task.
   Algorithm	Select an algorithm as required. Behavior analysis Target recognition Video structuring
   Video Type	Type of the video used for intelligent analysis. Select Live or Recording as required.
   Set Schedule	Set the time segment for executing the analysis task as required.

5. Configure the video structuring task, as shown in Figure 4-269.
   Figure 4-269 Configuring a video structuring task
   
   Table 4-243 Parameter description

   Parameter	Description
   Enable	Indicates whether to enable the current intelligent analysis task.
   Detection Area	Button for drawing a detection area. You can double-click the left mouse button to complete drawing. You can draw detection areas based on the following rules: You are advised to draw lines clockwise. The lines you draw cannot cross each other or form a sharp acute angle. Otherwise, detection areas may fail to be drawn. Do not draw polygonal areas that overlap with each other. Otherwise, false negatives may occur.

6. Click Save.

Definition

The device can search the snapshot library for similar objects based on personal attributes.

Customer Benefits

This feature allows users to search for data using a small number of conditions, improving search efficiency.

Application Scenario

This function is mainly used to configure target and vehicle recognition tasks in key areas such as campus roads and fire passages. The system can search for person information in the snapshot library based on personal or cyclist attributes.

Requirements

Application Limitations

This function is used only for PoC outside the Chinese mainland.

Scenario Description

Table 4-244 describes the support for this feature in different client scenarios.

Scenario Description

• Table 4-245 describes the support for this feature in different client scenarios.

• After local intelligence is enabled or an analysis task is created for a camera, do not modify the attributes of the camera. Otherwise, target recognition and video structuring tasks will become invalid, and you will need to delete and recreate the tasks.

Prerequisites

• The object classification or target recognition algorithm plug-in has been installed on the HWT-IVS1800 through the OMU portal. For details about how to install the plug-in, see Enabling Personal Feature Recognition in the HWT-IVS1800.
• Image storage parameters have been set. The detailed procedure is as follows: On the iClient S100 home page, choose Maintenance Management > Video Device > Device List. Right-click an HWT-IVS1800, choose Storage Management from the shortcut menu, select the desired camera, and set image storage parameters.

Procedure

1. On the iClient S100 home page, choose Maintenance Management > Video Device > Device List.
2. Right-click an HWT-IVS1800 and choose Task Management.
3. Click New Task.
4. Set video structuring global parameters, as shown in Figure 4-270.
   Figure 4-270 Setting video structuring global parameters
   

   Table 4-246 describes the parameters.

   Table 4-246 Parameter description

   Parameter	Description
   Camera	Select the camera for which you want to create an analysis task.
   Algorithm	Select an algorithm as required. Behavior analysis Target recognition Video structuring
   Video Type	Type of the video used for intelligent analysis. Select Live or Recording as required.
   Set Schedule	Set the time segment for executing the analysis task as required.

5. Configure the video structuring task, as shown in Figure 4-271.
   Figure 4-271 Configuring a video structuring task
   
   Table 4-247 Parameter description

   Parameter	Description
   Enable	Indicates whether to enable the current intelligent analysis task.
   Detection Area	Button for drawing a detection area. You can double-click the left mouse button to complete drawing. You can draw detection areas based on the following rules: You are advised to draw lines clockwise. The lines you draw cannot cross each other or form a sharp acute angle. Otherwise, detection areas may fail to be drawn. Do not draw polygonal areas that overlap with each other. Otherwise, false negatives may occur.

6. Click Save.

Procedure (on the LDU)

1. Log in to the LDU as the admin user. ( Logging In to the LDU)
2. Click in the upper left corner.
3. Choose Intelligent Application > Search > Person Search.
4. Configure person search, as shown in Figure 4-272.
   Figure 4-272 Configuring person search
   

5. View person search results.
   The system can search for images similar to the uploaded image. In the Search Results area, you can click a small image to view the corresponding large image for details, as shown in Figure 4-273.

   Figure 4-273 Search results
   

Procedure (on the iClient S100)

1. On the iClient S100 home page, choose Basic Operations > Search > Person Search.
2. Click the Filter by Criteria tab.
3. Set search criteria. Table 4-248 describes the key parameters.
   Table 4-248 Parameters for person search by criteria

   Parameter	Description
   Source	Cameras from which person images are to be searched for. You can select multiple cameras connected to multiple HWT-IVS1800s.
   Captured	Period in which person images are captured.
   Object Type	Type of the object for search, for example, Person or Cyclist. You can also set Tops Color, Backpack, Body Type, Helmet, Non-motorized Vehicle Color, and Non-motorized Vehicle Plate for accurate search.

4. Click Search. The Snapshot Library tab page with the search result is displayed by default, as shown in Figure 4-274. The snapshot library contains person images captured by cameras and extracted by personal feature recognition tasks or video structuring tasks.
   Figure 4-274 Person search result
   

5. (Optional) Use a clear person image for secondary search to obtain more images of the person. The procedure is as follows:

   Click a person image in the search result. On the details page that is displayed, click Search.

6. (Optional) Select multiple person images and click Export Selected to export the selected images. Alternatively, select Export This Page or Export All from the drop-down list box as required. Figure 4-275 shows the file export dialog box.
   Figure 4-275 Export Options dialog box
   
   

   You can enable the encrypted download function when exporting files. After this function is enabled, you need to enter the preset password when opening the file exported to the local computer.