Actually that's right you should use the enum.

Which OpenMV is that ? Also please include all the firmware info you see in the IDE when connected to the camera.

Yes but it's only available on the Arduino Portenta.

There's an example in the IDE and you can find the model here:

https://github.com/iabdalkader/microspeech-yesno-model

Copy it to SD card or flash storage.

BTW I also tested the firmware used in the linked post, it's fw 3.0.0, and I get the same 10mA with the OPENMV3/F7 in stop mode.

I just a made a small fix to shutdown and the final numbers that I see are 10mA in sleep()/lightsleep() (STOP mode), and less than 10mA in deepsleep() (STANDBY mode) I'm not sure what the exact numbers are if you have a meter hooked up let us know. 2. What do I need to modify in my code to reach 10m...

tomerle wrote: ↑

Mon Jan 25, 2021 9:55 am

So Currently it is not possible to get lower than 10mA on Sleep mode with the H7 plus ?

The best we can do right now is 30mA for STOP mode and less than 10mA in STANDBY mode. I'm working on a fix to the stop mode could possibly get down to 10mA.

Joe100A wrote: ↑

Sun Jan 24, 2021 3:53 pm

1) What is bpp below?
2) The seek function (seek(n)) hoes to the the nth frame and every time read() is called you get the n+1 frame out?

Bytes Per Pixel, and read returns the current frame not n+1 if you seek(0) read() returns the first frame.

Shields are stacked with the long headers, see the product page https://openmv.io/collections/shields/p ... i-shield-1

@Kwabena we shouldn't mention specific frequencies in the docs as it depends on the chip and board.

No, it must be used for the wifi module..

That's not my post and it's for the older camera not the H7+, actually the H7+ should be able to get to lower power or the same at least, since we fixed a missing pull up on the sensor power down.

Actually I realized QSPI wasn't disabled in low-power modes, after some fixes the board draws less than 10mA in deepsleep.. As for lightsleep, it still draws at least 30mA... However, I noticed if sensor.reset() is not called at all it gets down to 10mA.. so there must be some register that gets wri...

The lowest possible current is only achievable with the deepsleep example.

I get ~10mA with the deepsleep example (machine.deepsleep()) with the latest firmware.

There's no single answer that fits all, it depends on the flash FS that's used, on OpenMV boards without external SPI flash yes but only if you put the board in DFU manually, and then overwrite the filesystem page, on Portenta yes (the Arduino bootloader is DFU), on OpenMV boards with external SPI f...

I was thinking of getting the adafruit lsm9ds1 https://learn.adafruit.com/adafruit-lsm9ds1-accelerometer-plus-gyro-plus-magnetometer-9-dof-breakout as it should work with the same code as the accelerometer examples for the Nano33BleSense. Anyone know if this accelerometer works with the PortentaH7?...

The Portenta H7 doesn't have any IMUs as far as I know.

Try removing the SD and erasing the internal flash fs (there's a checkbox when uploading the firmware) also you must never use a different board firmware for yours.

No that's not possible.

I'll try to get a release out this month, but you should try to build and test this new feature so you can let us know if there's any issues.

Hi I implemented something that does exactly what you want, we didn't have this before because memory was low. After this is merged you should be able to record frames to memory and play them back and use the frames with any image processing function:

https://github.com/openmv/openmv/pull/1084

Hi, we use gcc version 7.3.1 20180622 (7-2018-q2-update) you can get it from here:

https://armkeil.blob.core.windows.net/d ... ux.tar.bz2

You just need to download the firmware zip from github releases and upload OPENMV*/firmware.bin for you camera, in this case it's OPENMV4. I don't think this feature will be added because we will have to include all the previous firmware releases and that will make the IDE much bigger, but feel free...

The bootloader times out how ? Can you post screenshots or a screen recording ? Also it might be worth trying on Linux, could be a Windows issue.

I think you're uploading the wrong firmware.

Hi, please try the latest firmware (3.8.0) first and then let us know if the issue is not fixed.

Thanks, I'll take a look. In the meantime, can you try setting the callback arg instead ?

Please report an issue on github, and make sure to fill the issue template.

It's possible to flash any firmware with the IDE, but it will not work after that with the IDE and you will not be able to flash a new firmware (without resetting the board manually). We don't support any firmware other than ours and we can't help you with development questions you have to know what...

Yes it's possible, the firmware should start at 0x08040000 and the bootloader will jump to it, you won't be able to load a new firmware without disconnecting/reconnecting the board.

Hi, there's no 3.9.6 firmware maybe you mean 3.6.9, in that case you should update to 3.8.0 the latest and then test again, if the issue is not fixed please report a bug on github with the smallest code that shows the issue.

OPENMV4.

people wearing masks or no masks from a side profile walking past an area that is properly lit with a plain wall background?

There's a model for front facing and H7+ only

https://github.com/SingTown/mask-tflite

The microcontroller has an RTC, however there's no backup battery, so you'll need to keep power on all the time. The RTC still functions in low-power modes. There's no plan to implement recording in memory for now, but if it ever gets implemented it will only be possible on the H7+ with SDRAM. EDIT ...

Actually I meant use EdgeImpulse to train a neural network, it can't train haar cascades, you can only do that with open-cv. If you already trained a cascade that works, just try to lower the threshold and/or increase the scaling step.

No, using smaller images ex 48x48, you don't have to retake photos I think the EI downscales the images.

> awsome thanks I'll adjust the sensor and retake the photos. You don't need to retain the network, just set the framesize to QQVGA (which is 160x120) so you should be able to scale down to 90x90, while saving some framebuffer RAM. Also retraining a smaller model or switching architecture should wor...

I see, the M4 core is not used right now in our firmware, but we may support it in the future... You can probably run vanilla/upstream micropython along with Arduino firmware, I'm not sure you should ask in the Arduino forums.

Yes to QQVGA or lower.

Wait what are you trying to do again ?

Try a smaller frame size, it will make more memory available to TF.

It's a lot easier to do classification with TF and EdgeImpulse and you'll get much better results. Please see the tutorials around just google openmv+edgeimpulse. Otherwise you should try tuning the threshold and scale factor, see the comments in the face detection example.

That indicates the flash cache is being flushed, there's no way to disable it right now it's hard-coded in micropython. If you use an SD it won't turn on any LEDs.

Does the command encompass the two core?

What do you mean?

Cool project! I don't see why not you may need to 3D print something to keep the cam in place but it should work.

Hi, not every module is available in micropython, please read the docs to see what you can import and how you can use it. This may be a good alternative, it's built-in you can just import it http://docs.micropython.org/en/latest/library/ubinascii.html Also this may or may not work, you need to downl...

In github releases, download the firmware.zip.

Yes, it gets collected when there's no free memory.

You should use the dfu-util that comes with Arduino IDE.