How to get and converge the face images from the cameras on the video surveillance management platform SVMSPro?

There are two solutions to get and converge the face images from the front-end cameras on the video surveillance management software SVMSPro.

Solution 1: If the front-end cameras already have face captured function and support GA/T1400 protocol, you can use video surveillance management software SVMSPro to manage those cameras directly, because the SVMSPro platform has already supported GA/T1400. All the face images can be transferred from cameras to the SVMSPro platform via GA/T1400 protocol to store and display.

How to set face recognition display via GA/T1400 protocol on SVMSPro?

A. Map the GA/T1400 face image transferring port.

The default port for face image transferred by GA/T1400 is 8203. You can find and modify this port in the sentence:

[GAT1400]

FacePort=8023

of the file “SConfig.ini”, which is under SVMS Pro server installation route, the default installation route is: C:\SVMS Pro\SConfig.ini .

Note: Please map this face default port 8023 before configuration. If not necessary, please don’t change this port.

B. Set view library (GA/T1400 protocol) parameters for front-end video devices. Set Hikvision and Dahua IP cameras for Examples.

1. How to set view library (GA/T1400 protocol)parameters for Hikvision IP cameras?

Procedure

Step 1: Log into Hikvision IP cameras’ website. Enable “Face Snapshot/Captured” menu, which you might find by going to Configuration > System Setup > Intelligent Resource Allocation. After setting, please click Save.

Step 2 : Go to Configuration > Network > Advanced Configuration > View Library, enable View Library. Please select view library version: VIID_201804, image uploading method: binary. Please input custom content for device ID.

Note: This device ID should not conflict with other devices’ ID. Please input admin for video library’s username and passwords. The server IP address is SVMSPro WAN IP address, input server ports: 8023.

Note: If you have changed the GA/T1400 face image transferring port, please input the changed port here correspondingly. Video channel’s ID is the same with device ID. After configuration, please click SAVE button.

Step 3: Click Face Snapshot/Captured to set corresponding face snapshot parameters.

Note: If using ONVIF protocol, default port is 80. If front-end camera has change the port, please input the port accordingly.

Note:
1). We only have Hikvision IP camera’s Chinese menu here, please find similar translation in their camera’s English UI.
2) Please make sure Hikvision IP camera has view library function and supports GA/T1400 protocol.

2. How to set view library (GA/T1400 protocol) parameters for Dahua IP cameras?

Procedure

Step 1: Log into Dahua IP cameras. Go to Setup>Network Setup > Platform Access > View Library.

Step 2: Enable the view library. The server IP address is SVMSPro WAN IP address, input server ports: 8023.

Note: If you have changed the GA/T1400 face image transferring port, please input the changed port here correspondingly. Please input any custom content for device ID.

Note: This device ID should not conflict with other devices’ ID. Video channel’s ID is the same with device ID. Select View Library Collection Objects you need. After configuration, please click Save button.

Step 3: Find Face Detection on the left menu, then enable this function. Input the corresponding parameters.

Note:
1) Face Detection must be enabled, if not, the captured face imaged cannot be transferred to platform.
2) We only have Dahua IP camera’s Chinese menu here, please find similar translation in their camera’s English UI.
3) Please make sure Dahua IP camera has view library function and supports GA/T1400 protocol.

C. Display face image on video surveillance management software: SVMSPro platform.

Procedure

Step 1: Run SVMSPro server and get SVMSPro WAN IP address.

Step 2: Set the IP camera’s view library, refers to B. Set view library (GA/T1400 protocol) parameters for front-end video devices.

Step 3: Run SVMSPro client program, and log into SVMSPro client. Go to Device Management > Face Recognition Device.Click Search, the Hikvision and Dahua IP camera will be displayed on SVMSPro. Select Device, then click Add to add this device. Input Device Name, then click Confirm to complete adding.

Step 4: Select one face snapshot camera on the left, then click Linked Channel to link the video camera for that face snapshot device.

Step 5: Click Face Recognition Display on SVMSPro, then select one linked video channel. Its video and captured face image will be displayed, see the image below.

Solution 2: If the front-end cameras don’t have face captured function and don’t support GA/T1400 protocol, you can install an AI box with face recognition function between camera and platform, using AI box to capture the face image and doing face comparison. Then transfer the face captured image and face comparison results from AI box to SVMSPro platform to save and display.

How to get the face image from AI box and display face recognition image and results on SVMSPro?

First of all, the video surveillance management software SVMSPro must have integrated AI box. The SVMSPro can integrate AI box by its SDK or any communication protocol, such as HTTP protocol.

Second set the face recognition parameters on AI box. Set RJ series AI box for example.

Procedure

Step 1: Log into RJ box web, with account name admin, and passwords admin. Go to Video Source > Add to add one video camera for AI box. The box supports ONVIF, RTSP, RTMP,HLS,EZV,DSS,ISC,GB/T28181 protocol, you can input camera’s corresponding video URL, such as RTSP URL, ONVIF URL etc.

Note: The image resolution for the added video should be 1080P.

After adding, the channel status will be displayed.

Step 2: Go to Face Database Management > Registry Face to add the face to the face database.

Face Id: Input any custom content for face ID

Note: Please don’t conflict with other face ID.

Name: Input the name for this face image

Save to Repository: You can create the new face repository by inputting the new name here. Or you can choose the current repository for this face image.

Face Image: Click + to add the face image for the face database.

Note: The size for face image <1MB, height <3900,width <3900.

Step 3: Add the AI algorithm task for that added camera video, by going to Task Management>Add

Task ID: Input the camera ID for the task ID.

Note: The task ID number should be same with video ID number.

Video Source: Select one video camera for the face recognition AI function.

Alarm Listener URL: Input SVMSPro platform IP address. The rule is: http://SVMSPro IP address:18886/aievent . For example: http://192.168.0.123:18886/aievent

Algorithm Settings: Enable Face Recognition.

Set Face recognition parameters below:

Threshold for similarity: compare the similarity with the face image in face database. The default value is 0.4. The higher value means more similar with the face picture in the face database.

Enable GA/T1400: Enable the face snapshot image to transfer via GA/T1400.

Enable Multi-Frame Judgement: When enable this selection, it will increase the demands for the environment, and it is hardly to detect the face image, when people are walking.

Threshold for strange: Input the threshold value for strange. If the similarity of the stranger is less than this value, it will be reported as stranger.

Alarm interval: Set alarm interval time to report the alarm information to the upper level platform.

Please select the face repository: Please select the face repository for the face comparison.

Minimum face pixel: Set the value for minimum face pixel. When the face pixel is lower than this defined value, it cannot detect the face image.

After configuration, please click Save to exit.

Step 4: Go to Option > Configuration Area to set the detection area for this video camera. After Setting, please trigger the alarm.

Step 5: Log into SVMSPro Client. Go to Device Management > AI server, the AI box will be auto registered and listed on the SVMSPro. Click AI Device on the left, then click Add To Management to add this AI box. Please don’t forget to input user name and password when adding.

Step 6: Select the AI box, then click Add Channel for this AI box. The Name and Channel No. should be the same with Task ID number which you set in the RJ series AI box.

Step 7: Selected the added channel, then click Linked Channel to associate the video camera for that added channel of AI box on SVMSPro.

Step 8: After configuration, please go to Control Panel > AI Alarm. Click one window, then drag the associated video channel to that window or double click the associated video channel under the AI box to open the video. When trigger the alarm, the face image will be transferred and displayed on the SVMSPro, see the image below. You can click Historical Record to check the face recognition alarm historical record information.