# Applicable specifications

OpenNCC DK, OpenNCC Knight ,OpenNCC Lite, OpenNCC USB

# Operation steps

# Preparatory work

One openncc camera, one USB 3.0 data cable with type-C interface, and one Ubuntu system computer; Connect the data cable to the USB 3.0 interface of the camera and the computer Figure-1

# Download the latest release package or source code of view

DownloadLink
If you start with source code,you need to Compile and run View,under CDK's QT project.

# Start from run View

# 1. OpenNCC operation permission

Enter CDK directory:Tools/deployment,run script “install_NCC_udev_rules.sh”,in the terminal,input:
./install_NCC_udev_rules.sh
After get permission to mount the OpenNCC,you need to restart the PC. Figure-2

# 2.Unzip the OpenNCC View packet,enter it,open a terminal,and input:

./AppRun,run the software(If can't run,please input:sudo ./AppRun try again)
After OpenNCC View opened,now we could using the View to work with the OpenNCC camera:
* Connect the OpenNCC Cam to the usb3.0 of the PC,clik Get device info button to get the device's information.If successful the information of the device would displayed on the Log area,If the connection is not 3.0, please rotate the data cable connected to the type-C interface of openncc camera by 180 ° and insert it again.Clik Get device info to get the device information. To make sure the log shows:USB interface is 3.0 ,and if shows not usb3.0,it also could be used as USB2.0. * Could choose one of the format from yuv420p/H.264/mjpeg,the default resolution is 1080p.
* Click to load model “1st network model”:,select a AI model.Currently already supported more than 10 models. If 'None' is selected means no AI model would be running in OpenNCC only streaming out the video.After you download a AI model,could use ROI function to select a area where you want the AI modle to recognize the scene in the region. * “Model Score” It is the lowest score of the algorithm recognition, and the algorithm will select the object in the frame only after reaching it;
* ”Display Scaler” is the video display window size, you can adjust the display window resolution.
* Check "show state" to select whether to display the current status information on the screen, including real-time frame rate, resolution and device ID.
* Check "inference accelerate" to select whether to enable algorithm acceleration. (must be selected before loading algorithm model)
* Click ”Start running models”,load the algorithm model and open the video stream successfully. Run the video stream, load the face detection algorithm, and then take a real shot: Figure-3

# 2.1 Demonstration of secondary model
  • Select vehicle-license-plate-detection-barrier-0106-fp16.blob from list “1st network model”
  • After selection, the required model can be selected under '2nd network model'. At present, only one secondary model (license plate recognition) is supported in the demonstration. Users can add more.

# Start with the SDK

  1. Download OpenNCC CDK,And install dependency package
  2. Enter Samples/How_to/Load a model
  3. make clean, make
  4. cd ../../bin/
  5. Connect the OpenNCC Cam with PC,and run ./Openncc

# Demo models of OpenNCC

Model Name Introduction Demo Application Other references
Object classification classification-fp16 ssd_mobilenet_v1_coco model can detect almost 90 objects Openncc Viewer OpenVINO Doc Link(opens new window)
Face and Person detection face-detection-adas-0001-fp16 Face detector for driver monitoring and similar scenarios. The network features a default MobileNet backbone that includes depth-wise convolutions to reduce the amount of computation for the 3x3 convolution block Openncc Viewer OpenVINO Doc Link(opens new window)
face-detection-retail-0004-fp16 Face detector based on SqueezeNet light (half-channels) as a backbone with a single SSD for indoor/outdoor scenes shot by a front-facing camera Openncc Viewer OpenVINO Doc Link(opens new window)
face-person-detection-retail-0002-fp16 This is a pedestrian detector based on backbone with hyper-feature + R-FCN for the Retail scenario Openncc Viewer OpenVINO Doc Link(opens new window)
person-detection-retail-0013-fp16 This is a pedestrian detector for the Retail scenario. It is based on MobileNetV2-like backbone that includes depth-wise convolutions to reduce the amount of computation for the 3x3 convolution block Openncc Viewer OpenVINO Doc Link(opens new window)
pedestrian-detection-adas-0002-fp16 Pedestrian detection network based on SSD framework with tuned MobileNet v1 as a feature extractor. Openncc Viewer OpenVINO Doc Link(opens new window)
People, Cars, Bicycles person-vehicle-bike-detection-crossroad-0078-fp16 Person/Vehicle/Bike detector is based on SSD detection architecture, RMNet backbone, and learnable image downscale block (like person-vehicle-bike-detection-crossroad-0066, but with extra pooling) Openncc Viewer OpenVINO Doc Link(opens new window)
pedestrian-and-vehicle-detector-adas-0001-fp16 Pedestrian and vehicle detection network based on MobileNet v1.0 + SSD. Openncc Viewer OpenVINO Doc Link(opens new window)
Vehicle Detection vehicle-detection-adas-0002-fp16 This is a vehicle detection network based on an SSD framework with tuned MobileNet v1 as a feature extractor. Openncc Viewer OpenVINO Doc Link(opens new window)
Mask Detect Mask-detect-fp16 Mask detect Openncc Viewer Under license
Vehicle License Plate vehicle-license-plate-detection-barrier-0106-fp16 This is a MobileNetV2 + SSD-based vehicle and (Chinese) license plate detector for the "Barrier" use case. Openncc Viewer OpenVINO Doc Link(opens new window)
Interactive face detection interactive_face_detection_demo This demo executes four parallel infer requests for the Age/Gender Recognition, Head Pose Estimation, Emotions Recognition, and Facial Landmarks Detection networks that run simultaneously Sample/Demo/work with OpenVINO/interactive_face_detection_demo OpenVINO Doc Link(opens new window)
Human Pose Estimation human-pose-estimation-0001-fp16 A multi-person 2D pose estimation network (based on the OpenPose approach) with tuned MobileNet v1 as a feature extractor. Sample/Demo/work with OpenVINO/human_pose_estimation_demo OpenVINO Doc Link(opens new window)