Hi! is it possible to assign a colour to a sound, so that when camera detects specific colour it playbacks assigned sound, and if it detects a number of colours at the same time all assigned sounds play simultaneously? Something similar to turntable sequencer but not in monochrome? I'm very new to electronics, sorry in advance if the question is not specific enough
Sir... I have a problem with our project.... When I put a rectangle within the coordinates and I want detect the color within the coordinates of where I draw the rectangle?? what find blob language will it use to detect a color in a specific coordinate area only?
@@OpenMVLLC Thanks, at first glance I thought it might be the PSVR project but then realized it is only doing 2D. I assume you are referring to the PHD sensor fusion stuff posted by Mariusz Wilk?
@@timlucas143 Yes, if you want 6 DoF I think you'd need more than 1 camera. Anyway, the answer to your question is basically yes, but, it depends on your skill level.
Hello again, can I print in the Serial Terminal an informations concerning to the different collors. For exemple "This is the reed/green color" or one number associated whith one combination of colors?
You should read the documentation. However, it's just selecting between which threshold to use for finding colors. docs.openmv.io/library/omv.image.html?highlight=find_blobs#image.image.find_blobs
My questions is what is " if blob.code() == 3: # r/g code" or "if blob.code() == 5: # r/b code" in the code? For exemple how can I detect exactly many different collors? Cause in the "if" part of the code is not very clear
Hi, please move questions to the forums. However, the documentation is quite clear on what that's for: docs.openmv.io/library/omv.image.html?highlight=blob#image.blob.code. code() is a bitmask of what colors made up the blob.
Please ask help support questions on the forum and provide enough details for an answer to be given. If you have non clone hardware then it would possibly be a brown out from a lack of USB voltage due to a bad cable or USB port on the computer.
Hi, I increased the resolution for B&W thus reducing the frame rate. B&W color tracking runs about 1.5x faster than RGB mode given the same resolution.
Hello OpenMV ! Super job on the improvements with the V7 ! Is there a way to grab at 60FPS in 640x480 ? (not talking about the tracking software capacity to process the data but more if the hardware is able to handle data at such a rate, enough RAM and CPU ?) Cheers, Jules
You could get the data into RAM at that rate, but, your not getting it out or processing it at that rate. Color Tracking at 320x240 runs at over 30 FPS now.
Let me expand on this. We can get a Bayer Pattern 640x480 image in RAM at 60 FPS. However, you can do anything in Python. You'll need to edit the firmware in C. You will only have time to process the image in between the lines that come in. So, basically, you'll want to capture each Bayer image line (2 lines likely) figure out the color of pixels, threshold them and then put the thresholded pixels into an output buffer. From there you can then find blobs on the binary image during the vsync period. If done right, you should be able to run at 60 FPS. But again, no python, no idea, just pure C.
Thank you for your super fast answer ! I will look into that as soon as I can order a new OpenMV :) I think I'll have to dig more into the code and I'll certainly have more questions then, but I'll start a thread on the forum I guess ;) Thanks a lot !
Would really appreciate better sound on the demo portion. Sound is great for the first section then drops to a blurred mumble with lots of echos for the second. Matching the audio levels & audio quality would improve the presentation. Content is great, though. Fantastic product!
It would be very interesting to compare it to Pixycam's color tracking performance. I use a Pixycam and it is very sensitive to lighting changes to the point that it makes it very difficult to use under any non-perfect lighting conditions.
We're kinda in the same boat. That's color tracking. However, we've got a nice color tracking threshold editor and stats package on the cam which will allow you to set very wide color thresholds so you don't have to worry too much. Grayscale tracking for example is a lot easier to do since you don't have to worry about color shifts. This means tracking lines and bright objects is really good. The big problem however is that the camera gain and white balance settings change each time you turn the system on. In the future we'll implement a function which allows you to easily backup and restore these settings given your color thresholds. However, this also will mean that the camera may not auto tune itself to the lighting in the room so you'll get different issues. In general, grayscale tracking works great even if the lighting is changing... and full color tracking works great as along as the lighting is constant.
We don't output sub pixel results. However, we are pixel accurate. What's cool is that we can actually emulate the pixy more or less. I'll be posting some scripts soon for this.
We don't have a lane detection algorithm on board. We have the components of it line finding lines however. But, not a complete easy to use function call for it.
Hi! is it possible to assign a colour to a sound, so that when camera detects specific colour it playbacks assigned sound, and if it detects a number of colours at the same time all assigned sounds play simultaneously? Something similar to turntable sequencer but not in monochrome? I'm very new to electronics, sorry in advance if the question is not specific enough
Yes, but you have to code that and we don’t offer sound playback services. So you will need to figure out how to make that work.
I want make a spectrophotometer.can you please give me any useful link for that ?
None of our customers have posted examples about that.
Sir... I have a problem with our project.... When I put a rectangle within the coordinates and I want detect the color within the coordinates of where I draw the rectangle?? what find blob language will it use to detect a color in a specific coordinate area only?
Hi, please ask for help support on the forums. In relation to your question use the "ROI" argument to specify a region of interest.
@OpenMV, LLC Could this be used to track LED's placed on an object to provide real-time outside in 6-DOF tracking of the object?
Yes, someone already did this. See our twitter feed. It's about a year back.
@@OpenMVLLC Thanks, at first glance I thought it might be the PSVR project but then realized it is only doing 2D. I assume you are referring to the PHD sensor fusion stuff posted by Mariusz Wilk?
@@timlucas143 Yes, if you want 6 DoF I think you'd need more than 1 camera. Anyway, the answer to your question is basically yes, but, it depends on your skill level.
how can I chose parameter Lmax, Lmin, A max, A min, B max B min. For example: Brown color
See the threshold editor in OpenMV IDE.
Have you ever put 2 openmv side by side and calculate the depth of an item being tracked?
You can do that with AprilTags tracking - it can tell you the tag X/Y/Z translation. But, anyway, no, I haven't don stereo vision.
Hello again, can I print in the Serial Terminal an informations concerning to the different collors. For exemple "This is the reed/green color" or one number associated whith one combination of colors?
Hi, the OpenMV Cam runs python scripts so you can make it print whatever you like. It's completely programmable.
Hi, another question. What does index in code mean? For example index = 2 for generic_blue_thresholds?
You should read the documentation. However, it's just selecting between which threshold to use for finding colors. docs.openmv.io/library/omv.image.html?highlight=find_blobs#image.image.find_blobs
My questions is what is " if blob.code() == 3: # r/g code" or "if blob.code() == 5: # r/b code" in the code? For exemple how can I detect exactly many different collors? Cause in the "if" part of the code is not very clear
Hi, please move questions to the forums. However, the documentation is quite clear on what that's for: docs.openmv.io/library/omv.image.html?highlight=blob#image.blob.code. code() is a bitmask of what colors made up the blob.
Sir... when I run my OpenMV and after a second, the connector will automatic disconnect... why is it??
O-Topia please ask for help support on the forum.
Please ask help support questions on the forum and provide enough details for an answer to be given. If you have non clone hardware then it would possibly be a brown out from a lack of USB voltage due to a bad cable or USB port on the computer.
What does it mean when appears treshold index not defined
It means you are indexing outside of an array.
sir... can we connect a button switch to openMV and interface it?
Yes, you can see the docs. Please ask these questions on the forums.
Curious why the color can do 30fps, but the B&W is only half that? I would have thought B&W would be more performant? Excited to try one of these!
Hi, I increased the resolution for B&W thus reducing the frame rate. B&W color tracking runs about 1.5x faster than RGB mode given the same resolution.
That makes sense, thanks for the reply
Hello OpenMV ! Super job on the improvements with the V7 ! Is there a way to grab at 60FPS in 640x480 ? (not talking about the tracking software capacity to process the data but more if the hardware is able to handle data at such a rate, enough RAM and CPU ?)
Cheers,
Jules
You could get the data into RAM at that rate, but, your not getting it out or processing it at that rate. Color Tracking at 320x240 runs at over 30 FPS now.
Let me expand on this. We can get a Bayer Pattern 640x480 image in RAM at 60 FPS. However, you can do anything in Python. You'll need to edit the firmware in C. You will only have time to process the image in between the lines that come in. So, basically, you'll want to capture each Bayer image line (2 lines likely) figure out the color of pixels, threshold them and then put the thresholded pixels into an output buffer. From there you can then find blobs on the binary image during the vsync period. If done right, you should be able to run at 60 FPS. But again, no python, no idea, just pure C.
Thank you for your super fast answer ! I will look into that as soon as I can order a new OpenMV :)
I think I'll have to dig more into the code and I'll certainly have more questions then, but I'll start a thread on the forum I guess ;)
Thanks a lot !
When i tried to run the code the While (true) is giving out an index error
Please ask for help support on the forums.
Would really appreciate better sound on the demo portion. Sound is great for the first section then drops to a blurred mumble with lots of echos for the second. Matching the audio levels & audio quality would improve the presentation. Content is great, though. Fantastic product!
You've got better editing skills than me. I did buy a copy of adobe premier. But, it's kinda laggy.
It would be very interesting to compare it to Pixycam's color tracking performance. I use a Pixycam and it is very sensitive to lighting changes to the point that it makes it very difficult to use under any non-perfect lighting conditions.
We're kinda in the same boat. That's color tracking. However, we've got a nice color tracking threshold editor and stats package on the cam which will allow you to set very wide color thresholds so you don't have to worry too much. Grayscale tracking for example is a lot easier to do since you don't have to worry about color shifts. This means tracking lines and bright objects is really good.
The big problem however is that the camera gain and white balance settings change each time you turn the system on. In the future we'll implement a function which allows you to easily backup and restore these settings given your color thresholds. However, this also will mean that the camera may not auto tune itself to the lighting in the room so you'll get different issues.
In general, grayscale tracking works great even if the lighting is changing... and full color tracking works great as along as the lighting is constant.
Does the tracking is outputting subpixels results ? Because as I remember the Pixycam just gave rounded pixel between 0 and 320.
We don't output sub pixel results. However, we are pixel accurate. What's cool is that we can actually emulate the pixy more or less. I'll be posting some scripts soon for this.
Are you using an openMV microphone?
The first part is from a cellphone mic and the second part is from a webcam mic. I'm not a professional video editor. I do what I can.
About 30 FPS for lane tracking, too?
We don't have a lane detection algorithm on board. We have the components of it line finding lines however. But, not a complete easy to use function call for it.
Sir why do we need to register for m7??
O-Topia please post questions on the OpenMV forum.
Where I can buy this in India ?
We don't have any Indian distributors. You have to buy internationally.
sir how to save prgramme in openmv m7 cam board
Use the IDE.
how to use face recognition process could you please make a video on that?
Please signup on the forums to ask questions.