What is machine vision
Machine vision is a comprehensive technology, including image processing, mechanical engineering technology, control, electric lighting, optical imaging, sensors, analog and digital video technology, computer software and hardware technology (image enhancement and analysis algorithms, image cards, I/O Card etc.). A typical machine vision application system includes image capture, light source system, image digitization module, digital image processing module, intelligent decision-making module and mechanical control execution module.
Application of machine vision in the industrial market
Using machine vision technology to replace labor can provide production efficiency and product quality. Therefore, machine vision technology is widely used in industrial inspection (measurement of the size and position of mechanical parts), robot vision, face recognition, license plate recognition, automatic optical inspection, and People drive cars, track and locate fields.
How to implement machine vision
Industrial machine vision systems include: light source, lens (fixed focus lens, zoom lens, telecentric lens, microscope lens), camera (including CCD camera and COMS camera), image processing unit (or image capture card), image processing software, Monitor, communication/input and output unit, etc.
The image processing unit, which can also be called a frame grabber, is a component of a complete machine vision system, but it plays a very important role. The frame grabber directly determines the camera interface (black and white, color, analog, digital), image processing, image output format, etc.
Tronlong's TLZ7x-EasyEVM evaluation board based on Xilinx Zynq-7000 SoC can well meet the image processing unit functions of industrial machine vision.
TLZ7x-EasyEVM evaluation board chip selection XC7Z020, compatible with XC7Z010, integrated PS-side single-core/dual-core Cortex-A9 ARM + PL-side Artix-7 architecture programmable logic resources, provides binocular camera interface, can be flexibly connected to the video output module.
Industrial machine vision-binocular image acquisition and processing example
1. Example function
Use Video In to AXI4-Stream IP core to collect data from 2 cameras (640*480@70), and cache it to PS side DDR through vdma, and then superimpose 2 channels of images into 1080P60 video through OSD IP core. Finally, display through VGA output.
2. The principle block diagram:
3. Example description:
This example uses the BlockDesign design method.
1) Camera video capture
Video capture uses the Video In to AXI4-Stream IP core, which is configured as Mono/Sensor in the routine, 1 pixels per clk, and each color data bit width is 8 bits. The routine uses 2 IP cores to collect 2 camera images respectively. The specific configuration is shown in the figure below:
2) Video data transmission buffer
Using VDMA (AXI Video Direct Memory Access) IP core, S2MM transfers the video stream to the DDR, and MM2S transfers the image data from the DDR. Example 2 VDMA IP cores, each IP core uses 4 frambuffer, the data width of stream is 8bits, as shown in the figure below:
3) Video stitching control
Use OSD (Video On Screen Display) IP core technology configuration. The OSD is configured with an AXI4-Lite interface, and the output resolution of the OSD, the number of superimposed layers, and the resolution and display position of each layer are set through the configuration register. This project is configured to 1080P resolution, 2 640*480 layers, without AXI4-Lite interface. As shown below:
4) Video display output (VGA)
Use AXI-Stream to video out IP core to display video, configure its video format as RGB, 1 pixels per clk, and each color data bit width is 8bits. The specific configuration of the IP core is as follows:
The detailed user manual, supporting module information, case source code, etc. of this example can be downloaded by clicking this link: http://site.tronlong.com/pfdownload
4. Hardware connection:
5. Operation effect:
6. Case video:
In this framework, users can perform richer configuration processing on video images according to their needs, such as edge detection, which can be applied to binocular stereo vision, virtual reality and other occasions.