The familiar word of vision has appeared more and more in front of the public. With the development of computer technology and the advancement of industry to intelligence, more and more visual recognition has been applied to all walks of life. The images that were once taken from the camera show us a lot of information. Driven by current smart equipment and technological innovation, the camera brings us more things. Automated detection is an inevitable trend in place of labor. Various industries are emerging in a variety of applications, such as medical image recognition, object detection on production lines, and warehouse material identification.
Computer vision, or machine vision, acquires image information from a camera and recognizes objects from the image to achieve the purpose of simulating the human eye to determine what the object sees.
1. Support multiple cameras
At present, Atom Vision software supports Hikvision cameras, Basler cameras, and Dahua cameras. Common settings such as exposure time, trigger mode, and debounce time can be set for the connected camera.
2. Contains multiple recognition algorithms
Includes recognition of circles, object size and angle recognition, and image template matching.
It can achieve sub-pixel level recognition accuracy and a recognition speed of several milliseconds to meet most application scenarios, providing precise and fast target positioning for mechanical actuators.
3. Result output
After visually recognizing the object, you can output the object coordinate, angle and other identification information through tcp communication, or let the output IO value to the relay.
Multiple output filters and conversions ensure that the output is the type of mechanical structure desired.
4. Integrate multiple tools
Load the image folder to identify the image in the folder, and it is convenient to test the visual recognition effect and adjust the recognition parameters without connecting the camera.
Measure the length and angle of the line segment in the image, measure the length, width, area and angle information of the rectangle to facilitate the setting of the identification parameters.
Flexible switching of recognition objects makes switching between different recognition scenes quick and easy.