On July 5, 2020, VisionChina 2020 -- China (Shanghai) Machine Vision Exhibition and Seminar on Machine Vision Technology and Industrial Application successfully concluded in National Convention and Exhibition Center. Suzhou Institute of Automation, Chinese Academy of Sciences (CAS) brought a series of hardware and software products, including a deep learning visual analysis platform, 3D structured light monocular camera, 3D structured light binocular camera, 3D linear laser camera, 2D ultra-high speed camera, and embedded AI intelligent camera, to the booth at 7.1F123, demonstrating prospective application technology research and industrialization work. For China's high-end intelligent manufacturing to provide core, basic technology platform and software and hardware integration solutions.
As a machine vision exhibition integrating scale, professionalism and authority in China, VisionChina (Shanghai) will continue to expand to vertical industries and technologies in addition to consolidating the large, comprehensive and sophisticated products on display. From innovative technologies to market trends, from application solutions to successful cases, it provides the most authoritative communication platform for the audience and participants to understand the development of the entire machine vision industry, attracting nearly 3,000 exhibitors, the exhibition scale of 160,000 square meters.
Deep learning visual analysis platform "Tianduan"
Tianduan 2020 version interface
Since the day break, Sue institute division line, intelligence analysis platform contains deep learning vision image annotation, training and inference of three main modules, can carry on the classification to the image, segmentation and positioning of the target function, the input image is compatible with 2 d black-and-white and color digital image, for a variety of format images, suitable for industrial flaw detection, localization scenarios such as object recognition and classification and recognition. "Tianduan", a deep learning visual analysis platform, is released by the Suzhou Institute of Science and Technology.
3D structured light monocular series cameras
Since the Soviet union in the institute, the division in line wisdom light series of monocular camera 3 d structure is based on structure light measurement method of non-contact measuring instrument, the main approach to realize the object to be measured for projection equipment project information encoded with stripe light, with light cameras record the stripe image sequence, the use of specific algorithms, it is concluded that the 3 d data of the object to be tested. The hybrid structured light technology based on phase-shifting method and Gray code can complete 3D reconstruction of the whole surface without moving the camera or the measured object, which not only saves the cost of high-precision mobile equipment, but also makes installation, debugging and maintenance more convenient. The first generation of 3D structured light camera that can reconstruct 3D point cloud in real time online has been released by the Chinese Academy of Sciences and the Chinese Academy of Sciences
3D structured light binocular camera
China - Suzhou Research Institute · China Kexingzhi 3D structured light dual camera
Product features
High frame rate: 3fps
Ultra-high accuracy: Z-axis repetition accuracy is 1μm, absolute accuracy ≦10μm
3D point cloud post-processing: automatic stitching and measurement of 3D point cloud
Support for CUDA acceleration: faster point cloud reconstruction
A variety of specifications for different environments: can be large field of vision, high precision, can also be small field of vision, ultra-high precision, flexible to deal with a variety of scenarios
Industrial protection: not afraid of harsh industrial application environment
Product parameters
Click on the image to view a larger HD image
3D line laser camera
China - Suzhou Research Institute · China Science Xingzhi 3D line laser camera
Product features
Out of the box: pre-calibration, calibration files are burned in the camera, out of the box
High Speed: Scan speed up to 6.4KHz
Real-time: Built-in FPGA preprocessing, real-time transfer of gravity center data
High protection grade: industrial IP66 aluminum alloy shell, integrated molding, can meet all kinds of harsh working conditions
High precision: Z-axis repeatability is up to 0.4μm, which can detect all kinds of micron-level high precision parts
Application scenarios
2D ultra-high speed camera
2D ultra - high speed camera of China - Jiangsu Institute of Research
China Self-Jiangsu Research Institute · China Kexingzhi 2D ultra-high speed camera is the first self-owned brand in China that supports online real-time processing of high-speed vision. It can monitor high-speed moving objects in real time, analyze images in real time, dynamically save the time period when problems happen and conduct follow-up analysis.
Product features
High-speed online real-time processing: real-time distortion correction, counting, centroid calculation, defect detection
Ultra high speed: 1280x180@18000fps
Operating under harsh working environment: protection class IP66, anti-corrosion class WF2, -40℃~45℃
Support secondary development: provide SDK to support secondary development of customers
Sufficient storage: 16GB, 3-second high-speed video, 15,000 frames of full-frame image
Application scenarios
Embedded AI smart camera
China - Suzhou Research Institute · China Kexingzhi embedded AI intelligent camera
Embedded AI intelligent camera has a built-in efficient operating system, which can be used independently from IPC. Based on ARM architecture and using FPGA acceleration algorithm, compared with the traditional camera + IPC mode, the operation is stable and efficient. Integration of self-developed GIVS visual algorithm platform, can independently complete positioning, identification, measurement, code reading and OCR and other conventional visual detection tasks. In addition, the AI deep learning function module is added to greatly improve the detection rate of weak defects or complex background targets, making the smart camera capable of more complex detection scenes.
Product features
System architecture: based on ARM+AI chip, high speed and stable operation, support FPGA acceleration
Flexible control development: USB direct keyboard and mouse, remote control, terminal command control
Multi-function algorithm module: positioning guidance, measurement, OCR, code reading, defect detection
AI Feature: Integrated deep learning neural network
Interface rich: GPIO, RS232, Ethernet, HDMI
Trigger mode: single frame trigger, multi frame trigger, software and hardware trigger
Other Features: Customized configuration of user indicator, support for log saving and export, IP65 protection, optional light module
Application scenarios
More Amazing Moments
