Framework for supporting other CMOS OR CCD Sensor

Hi,

Do you guys plan to have some kind of framework .Structure i.e your wishbone type structure to support other type of CMOS sensor or CCD sensor . it would be interesting to have this . it will help integrate othertype of sensor on this board. or you can lay out format then people will follow up the same of the implementation.

We were thinking to interface Sony line scan sensor with your board.

Comments

  • Hi,

    thanks for your interest. The architecture we use for the sensor is already quite general right now. What we have is a component that provide the interface with the sensor and then another component that takes data from the sensor and format it to be written into a FIFO. The FIFO is then connected on the wishbone bus and thus accessible from the processor. The only problem we have right now is that the FIFO only uses integrated block ram (we use an 8KB FIFO) and depending on image/data scale and rate we can have problem preventing the FIFO to overflow.
    I'am trying to get all this integrated into skeleton so people can use the camera with no need to write HDL. Its not finalized yet but it should be in near future (less than a month). I'll keep you updated on this.

    The line scan sensor seems a nice item to interface with especially to design  LIDAR like applications. Do you have any reference to the line scan you plan to use ? I 'am currently building an application with a laser line and the ov7670 to detect obstacle using video.

    Regards,
    Jonathan Piat
  • Regarding FIFO with on board SDRAM . i have one USB3 Kit from Haunor which has Cyclon III and has Two SD ram. htey have written Hardware RAM access controller for FIFO  write to FX3 and read FIF from FX3 32 bit . if you want the VHDL code for that i can send in that you . that might be of help to implement FIFO through SDRAM.

    OV7670 try to use this in linescan mode scan only two or 4 lines . at highest possible rates . i.e at 9K hz would be fine . you can also realize have laser triangulation sensor with that if you can implement Triangulation calculation . it will be 3D line laser scanner !!!
  • I have just finished a first step with the ov7670 where i use it at 320x240 (QVGA) mode and use a laser line. I filter the image through sobel and detect the max vertical gradient in each column in the image to detect the laser line. Then by comparing the line position in each column to the expected position i can detect obstacle and i could triangulate their 2D position if i had a well calibrated system (it would require to calibrate the camera and laser).
    Its a very good idea to use only 4lines of the sensor to use it as a linescan ! I need to verify that i can configure to such a low res, but thoertically it could work and lead to a cheap LIDAR implementation.

    I'am very interested by the SDRAM fifo code, this is what currently limit the processing to QVGA if i want to be able to capture an image. VGA would allow me to calibrate teh system and have better results. Please send the code to jpiat@valentfx.com.

    Thanks,
  • ok i will E-mail the complete code to you
  • Have u got chance to review fifo vhdl code i sent u. It is for altera but design should be same for xilinx .
  • i downloaded the code but i did not have a chance to look more into it. The example code is scattered into multiple times, so it may take time to understand and port to logi.
  • could you define block diagram for SDRAM from code that i sent you. i can try to design SDRAM code from my friends.
  • i found this link has designed http://hamsterworks.co.nz/mediawiki/index.php/Simple_SDRAM_Controller  SDRAM controller for Logi Bone . is this working can we use it .
  • This is already what we use for our tests of the SDRAM. I have an SDRAM fifo based on this controller almost ready.
  • Hi,

    Any progress on the SDRAM interfacing front . we are eagerly waiting for this .

    We are planning  smart camera plateform based on this . like www.microscan.com . We are preparing software based on .net and will run it using .net mono plateform. User interface will be based on .net windows forms . and all image processing will be done using Opencv.

    if SDRAM interface works then we can process bigger images.
Sign In or Register to comment.