Smart Vision

Intelligent computer vision system For smart factory, smart farm and smart city
through OpenCV and Deep Learning technique in
low cost embedded device.

Deep Learning-based Intelligent Marchine Vision Solution

Smart Vision which was developed by OPENSYSNET’s own technology is intelligent computer vision solution
that can be applied various industry such as smart factory, smart farm, smart city.

SMART VISION can be used to

Smart Vision is an intelligent solution that can be work quickly and accurately instead of human.
It can be used in various fields like defect detection, component alignment as a part of smart factory.

Defect Detection Vision Solution is a solution that can improves the quality of manufacturing process and
productivity by determining defected product and minimizing defect rate though vision monitoring in the factory manufacturing line

    Pre-Learning for Products and Defect cases
    Though pre-deep learning, Vision Solution determine
    whether product defectiveness automatically.
    Product Inspection
    Product inspection and defect detection
    through deep learning algorithm
    Management Specification
    Though pre-deep learning for product specification,
    Vision Solution can manage specification of each product.
    Statistics Management
    Enhancement product quality and minimizing
    defect rate through managing inspection result and statistics.

Advantages of Defect Detection Vision Solution

Our smart vision is a video analysis and detection solution based on deep learning.
It also provides high performance of defect detection and high-level user convenience.
    Pre-learning and Inspection based on
    Deep Learning
    Can detect defectiveness precisely based on
    deep learning training for various defect cases.
    Excellent Defect Detection and Fast Inspection
    High speed inspection based on training data and
    do not need any supervisor for monitoring because of
    our statistics management function
    Support User-Friendly UI
    Support convenience UI that user can do everything
    such as system setup, defect detection, statistics
    management and specification management.

Reference for Defect Detection Vision Solution

Our solution has been verified its performance in PCB defect detection system which can detect open circuit,
checking component position, detecting soldering on PCB

    User Login
    User Registration
    Can add/delete user and mange user’s
    authority through User setup menu
    System Setup
    Through system setup, User can change
    any setting if any peripheral devices
    such as camera was changed
    Recognition PCB
    Can setup a PCB to inspect easily
    and simply through our web service
    Setup Detection Type
    Can setup inspection type
    among various detection type
    Inquire Defect Detection
    Just one click in detection history menu,
    user can check defectiveness info such as
    location of defect, defect item, etc.

Defect Detection Vision Solution can be used in

Our vision solution can be applied in various industries such as containers manufacturing, inspection packing status, inspection product appearance, etc.
And customer can experience decreasing defect rate and cost cutting for dealing with claim through our high performance AI solution

    Semiconductor Component
    Can inspect any defectiveness such as
    the presence or absence of parts, mis-insertion, etc.
    Inspection Packing state
    Can inspect various cases such as
    crooked label, packing defect, barcode, etc.
    Apperience Inspection
    Can inspect product appearance
    status, expiration date, etc.
    Manufacturing Line
    Any components can be inspected quickly
    through AI detection solution
    Inspection PCB circuit line
    Inspect any circuit line on PCB

OPENSYSNET Vision S/W is performing for detecting parts and calculating angle according to
the pre-defined shape to interworking with alignment robot in factory automation

  • Smart Vision
  • Detecting Parts
    • Can be detecting parts on the feeder by using vision algorithm according to configuration target part information.
    • Can calculate the location information of the detected parts through calibration.
    Calculating Angle
    • Can calculate angle for detected parts according the feature information.
    • Can config the unique feature information of the part by using the management web UI/UX.
    • Can be worked with various H/W such as vision camera, flex feeder and robot for alignment operation.
    • Can be used for fault detection in assembly line in factory automation.
    Management Web
    • Provides Web based management UI/UX for configuration and monitoring in real-time
    • Consists of working area, part configuration, functional test and monitoring.
  • Smart Vision
    Web Login / System Info
    User can access the system management by using proper account and check system information such as serial number, IP address, etc.
  • Smart Vision
    Setting Working Area and Target Part
    Setting working area on feeder and basic information of parts.
  • Smart Vision
    Functional Test & Real-time Monitoring
    Can check the detection accuracy of solution through functional test function and can monitor the robot working process through real-time monitoring function.

Our Smart Vision Solution is collaborating with SCARA robot for harvesting tomato, pepper, strawberry as smart farm 4.0 project.

Use Case

Smart vision box can train and detect any kind of crops based on Baseline Learning and Transfer Learning of deep learning detection model.

Crops Harvesting Vision Solution

    Building various training data through
    web auto crawling and data augmentation
    Building training data set through auto
    web image data collection and data augmentation
    Support various robot through
    own deep learning edge box
    Deep Learning edge box with LiDAR camera
    and NVIDIA Jetson NANO module
    Adapt various crops through
    deep learning training
    Based on training data of crops to be harvested,
    can apply various crops through deep learning training
    Support various robot through
    own deep learning edge box
    Deep Learning edge box with LiDAR camera
    and NVIDIA Jetson NANO module
    Max view angle 70, 55 degree
    and max 23.6M distance
    Max 23.6M distance measurement based on
    1024*768 resolution of LiDAR camera
    Support robot control function
    Robot movement and location control according to
    video analysis of harvesting crops

Smart Vison Box For Smart Farm

OPENSYSNET Smart Vision Box is edge type AI vision box which perform object detection and recognition rapidly
based on LiDAR camera and Deep learning object detection model.

  • Smart Vison Box (Front)
    Smart Vison Box (Back)
  • Item Specification
    CPU Quad-core ARM Cortex-A57 MPCore processor
    GPU NVIDIA Maxwell architecture with 128 NVIDIA CUDA® cores
    Memory 4 GB 64-bit LPDDR4, 1600MHz 25.6 GB/s
    Storage 16 GB eMMC 5.1
    Video Encode 250MP/sec
    Video Decode 500MP/sec
    Boot Source eMMC / USB
    Networking Gigabit Ethernet, M.2 Key E
    Peripheral Interface 4x USB 3.0, USB 2.0 Micro-B
    HDMI 2.0 and eDP 1.4
    LiDAR Camera Ideal Range : 0.25m ~ 9m
    FOV : 70º * 43º
    Accuracy : 5mm ~ 14mm
    Frame Rate : 30fps


Our software can be performing various type of object detection and recognition based on AI deep learning training.

  • Competitiveness HIGH PERFORMANCE

    High-Accuracy & High-Speed Detection

  • Competitiveness INTER OPERABILITY

    Working with Various Robot & Camera

  • Competitiveness COST EFFICENCY

    Low Cost for Installing S/W with Camera

  • Competitiveness SUPPORT MULTI-PLATFORM

    Support various platforms such as CPU, GPU and TPU

Deep Learning-based Machine Vision Solution

Machin Vision Solution based on deep learning supports intelligent vision service that provide human’s vision and judgment ability .

Contact Us

Please feel free to contact us if there is any questions.
We will get back to you soon.

위로 가기