implement own Haar-Cascade

[fabianhummel]

Hi,

currently I am trying to implement my own hardhat-recognition to my MV-cam. For the Cascade classifier I used openCV and with my webcam it is working. Then I created a new .xml-classifier-file with the additional parameters "-SaveBaseFormat" and "-FeatureType HAAR" in order to make it usable for the MV-cam. Then I used this (https://github.com/openmv/openmv/blob/m ... cascade.py) script to transform it into a .cascade-file which I included like this into my micropython sript:

# the cascade script is on the internal flash drive
hat_cascade = image.HaarCascade("/cascade_scenario4.cascade", stages=15)

...

while (True):
clock.tick()

# Capture snapshot
img = sensor.snapshot()

# Find objects.
# Note: Lower scale factor scales-down the image more and detects smaller objects.
# Higher threshold results in a higher detection rate, with more false positives.
objects = img.find_features(hat_cascade, threshold=1, scale_factor=1.35)

# Draw objects
for r in objects:
img.draw_rectangle(r)


But my script just does nothing. It gets the snapshots, but recognized nothing. Does anyone have any experience in including own haar-cascades?
Did I implement it right?

best regards

[kwagyeman]

I've altered Ibrahim to this issue. Not sure how to start the debugging process for this as I don't know this part of the code.

[kwagyeman]

Can you print the Haar Cascade object to see if it was loaded without issues?

Code: Select all

print hat_cascade

[iabdalkader]

Hi, I'm not sure why it's not working, try to remove the stages argument to test all stages. If you can you attach the xml and binary cascades I'll test them for you.

[fabianhummel]

Hi

yes the print(hat_cascade) works. It prints values like:
20.13845
20.13423
20.13026
20.12651
20.12297
20.11963
...

I attached the XML, but I can't attach the .cascade-file due to the filetype..?

Best regards

[kwagyeman]

Zip it up.

[iabdalkader]

It seems to be working on some hats and not others:



Note it has 16 stages (the script prints cascade info) you should use them all.

Update:
I used all cascade stages by just leaving out the stages arg:

Code: Select all

hat_cascade = image.HaarCascade("/cascade_scenario4.cascade")

Also used a 0.85 threshold and 1.15-1.25 scaling factor.

[fabianhummel]

Thanks for the replies
In the attachment ist the zipped up XML & Cascade

best regards

[fabianhummel]

Sorry for asking again, but can someone tell me wheather this .cascade-file is working?

Currently, when I run this code:

import sensor, time, image, pyb

# Reset sensor
sensor.reset()

# Sensor settings
sensor.set_contrast(1)
sensor.set_gainceiling(16)
# HQVGA and GRAYSCALE are the best for face tracking.
sensor.set_framesize(sensor.HQVGA)
sensor.set_pixformat(sensor.GRAYSCALE)

greenLed = pyb.LED(2) # define green LED.

# Load Haar Cascade
# By default this will use all stages, lower satges is faster but less accurate.
hat_cascade = image.HaarCascade("/cascadesc22.cascade")
print(hat_cascade)
# frontalface
# /cascade_scenario4.cascade
# FPS clock
clock = time.clock()

while (True):
clock.tick()

# Capture snapshot
img = sensor.snapshot()

# Find objects.
# Note: Lower scale factor scales-down the image more and detects smaller objects.
# Higher threshold results in a higher detection rate, with more false positives.
objects = img.find_features(hat_cascade, threshold=0.85, scale_factor=1.2)

# Draw objects
for r in objects:
img.draw_rectangle(r)

greenLed.on()
time.sleep(10)
greenLed.off()

# Print FPS.
# Note: Actual FPS is higher, streaming the FB makes it slower.
print(clock.fps())


I just get the line: width:36 height:36 n_stages:17 n_features:150 n_rectangles:260
And the red-LED turns on and I can't do anything, not even disconnecting :/

What am I doing wrong?
attached the cascade-file

best reards

[kwagyeman]

Hi, I've asked Ibrahim for help.

[iabdalkader]

No this one doesn't work, when I run the script my cam hangs, if you can attach the xml file for this cascade I'll regenerate it and test again.

[fabianhummel]

Hi,

I attached the XML-file.
Hope this helps solving my problem

best regards

[iabdalkader]

Hi,

I attached the XML-file.
Hope this helps solving my problem

best regards

Hi,

This one works (doesn't crash the camera) it's different from the one you sent before (cascadesc22.cascade) I'm attaching the binary cascade I made with the script

[fabianhummel]

Hi all,

thanks for the replies!

I have another question:
I created the XML-file with openCV. In order to run it on my Laptop (using it's webcam) I created the new version of the XML-file (without the parameter -SaveBaseFormat).
The result is very good an it detects a lot of hardhats.
Now, in order to run it on the openMV-Cam I need the old format of the XML-file (created by using the parameter -SaveBaseFormat). But then I use the script (https://github.com/openmv/openmv/blob/m ... cascade.py) in order to convert it into a .cascade-file and it detects almost nothing. How ist this possible: I used the same parameters as I did for the new format. :/

Is this a common issue=?

best regards

Fabian

[kwagyeman]

Have you tired playing with the threshold and scale ratio?

[Charlie]

Hi iabdalkader,
I tried to replace the eye detection cascade with a new one. I converted the .xml file to .cascade file using the provided tool. But the video stream stuck and I think the camera cracked . I read the thread above and bascially the problem is the same as Fabianhummel's. You created a new binary cascade (cascade.zip) for him and it also works on my machine. I want to know how you did that. I also attached my xml file. It'll be great help if you create a binary cascade for me.

Best regards,

[iabdalkader]

Hi, it seems the cascade tool only works on Linux.

[SADEGHRAHMANI1374]

Hi, it seems the cascade tool only works on Linux.

hi @iabdalkader
i want to create a cascade file from xml file
xml file is a opencv sample (haarcascade_fullbody)
i try to convert xml file to cscsde file whit openmv-cascade.py but it return


usage: conver [-h] [-i] [-n NAME] [-s STAGES] [-c] file
conver: error: too few arguments

Traceback (most recent call last):
File "C:\Users\azaduni\Desktop\conver", line 288, in <module>
main()
File "C:\Users\azaduni\Desktop\conver", line 272, in main
args = parser.parse_args()
File "C:\Python27\lib\argparse.py", line 1656, in parse_args
args, argv = self.parse_known_args(args, namespace)
File "C:\Python27\lib\argparse.py", line 1688, in parse_known_args
return self._parse_known_args(args, namespace)
File "C:\Python27\lib\argparse.py", line 1901, in _parse_known_args
self.error(_('too few arguments'))
File "C:\Python27\lib\argparse.py", line 2311, in error
self.exit(2, _('%s: error: %s\n') % (self.prog, message))
File "C:\Python27\lib\argparse.py", line 2299, in exit
_sys.exit(status)
SystemExit: 2

its my xml file
and its openmv-cascade.py whit path chenged

[iabdalkader]

I don't know what conver.py is, please use our script:

./openmv-cascade.py haarcascade_fullbody.xml

Also Note it doesn't work on Windows.

[SADEGHRAHMANI1374]

I don't know what conver.py is, please use our script:

./openmv-cascade.py haarcascade_fullbody.xml

Also Note it doesn't work on Windows.

thank you for answer to my question
i use your script and now had this problem

[kwagyeman]

Please post the steps of what exactly you did.

[SADEGHRAHMANI1374]

Please post the steps of what exactly you did.

i download openmv-cascade.py and insert path of xml file in script
when i want run script in raspbian return this error

[iabdalkader]

Can you run it from the terminal so we can see the error ?

[iabdalkader]

here I converted the file for you.

[SADEGHRAHMANI1374]

here I converted the file for you.

Is there any tutorial to learn this topic like this??



https://www.youtube.com/watch?v=I8s4PDrIV4E

[iabdalkader]

No the script should work out of the box though, anyway we're going to add a built-in cascade converter to the IDE it's just a matter of time.

[SADEGHRAHMANI1374]

Can you run it from the terminal so we can see the error ?

[kwagyeman]

It looks like you have html in your script... i.e. you didn't download it right...

[SADEGHRAHMANI1374]

It looks like you have html in your script... i.e. you didn't download it right...

[SADEGHRAHMANI1374]

It looks like you have html in your script... i.e. you didn't download it right...

thanks for replying

How to write the xml file to the script or in the other words how to Introduce xml file to script?

[kwagyeman]

Hi, you pass the xml file after the script name.

[SADEGHRAHMANI1374]

Hi, you pass the xml file after the script name.

I put the xml name in the script or in the terminal ?
if i must put in script
It is possible to place the modified file?

[iabdalkader]

Hi download the script like so:

Code: Select all

wget https://raw.githubusercontent.com/openmv/openmv/master/usr/openmv-cascade.py

Then run it like this:

Code: Select all

python2.7 openmv-cascade.py cascade_file.xml

[SADEGHRAHMANI1374]

Hi download the script like so:

Code: Select all

wget https://raw.githubusercontent.com/openmv/openmv/master/usr/openmv-cascade.py

Then run it like this:

Code: Select all

python2.7 openmv-cascade.py cascade_file.xml

Thank you so much

[aqeelyaacob]

hai,i got a problem here,when i try to convert the xml file using converter as here : https://github.com/openmv/openmv/blob/m ... cascade.py
i got an error. So here i attach the xml file.Seem i stuck to convert to cascade file.After that,does the cascade embedded directly to camera or i need to transfer the cascade into internal memory to load on another script.

[iabdalkader]

hai,i got a problem here,when i try to convert the xml file using converter as here : https://github.com/openmv/openmv/blob/m ... cascade.py
i got an error. So here i attach the xml file.Seem i stuck to convert to cascade file.After that,does the cascade embedded directly to camera or i need to transfer the cascade into internal memory to load on another script.

The converter doesn't work on Windows, try the attached cascade.

You need to copy the cascade to flash or sd card, see examples and replies to this post.

[aqeelyaacob]

hai,i got a problem here,when i try to convert the xml file using converter as here : https://github.com/openmv/openmv/blob/m ... cascade.py
i got an error. So here i attach the xml file.Seem i stuck to convert to cascade file.After that,does the cascade embedded directly to camera or i need to transfer the cascade into internal memory to load on another script.

The converter doesn't work on Windows, try the attached cascade.

You need to copy the cascade to flash or sd card, see examples and replies to this post.

it all done,thanks to you.But the detection still low rate,does i need to add more positive image during making a cascade?.here i attach the video and also coding



import sensor, time, image

lens_mm = 2.8 # Standard Lens.
average_spool_height_mm = 360.0 #
image_height_pixels = 240.0 # QVGA
sensor_h_mm = 2.952 # For OV7725 sensor - see datasheet.
offest_mm = 100.0 # Offset fix...

# https://photo.stackexchange.com/questio ... in-a-photo
def rect_size_to_distance(r): # r == (x, y, w, h) -> r[3] == h
return ((lens_mm * average_spool_height_mm * image_height_pixels) / (r[3] * sensor_h_mm)) - offest_mm

# Reset sensor
sensor.reset()

# Sensor settings
sensor.set_contrast(1)
sensor.set_gainceiling(16)
# HQVGA and GRAYSCALE are the best for face tracking.
sensor.set_framesize(sensor.QVGA)
sensor.set_pixformat(sensor.GRAYSCALE)

# Load Haar Cascade
# By default this will use all stages, lower satges is faster but less accurate.
spool_cascade = image.HaarCascade("myhaar.cascade", stages=1)
print(spool_cascade)

# FPS clock
clock = time.clock()

while (True):
clock.tick()

# Capture snapshot
img = sensor.snapshot()

# Find objects.
# Note: Lower scale factor scales-down the image more and detects smaller objects.
# Higher threshold results in a higher detection rate, with more false positives.
objects = img.find_features(spool_cascade, threshold=0.75, scale_factor=1.25)

# Draw objects
for i in range(len(objects)):
img.draw_rectangle(objects)
img.draw_string(objects[0] - 16, objects[1] - 16, "Distance %d mm" % rect_size_to_distance(objects))

img.draw_cross(img.width()//2, img.height()//2, size = min(img.width()//5, img.height()//5))

# Print FPS.
# Note: Actual FPS is higher, streaming the FB makes it slower.
print("FPS %f" % clock.fps())

[iabdalkader]

it all done,thanks to you.But the detection still low rate,does i need to add more positive image during making a cascade?.here i attach the video and also coding

I'm not sure, maybe try a lower scaling factor (1.15).

[aqeelyaacob]

it all done,thanks to you.But the detection still low rate,does i need to add more positive image during making a cascade?.here i attach the video and also coding

I'm not sure, maybe try a lower scaling factor (1.15).

it's done.thanks..Now i'm designing alignment algorithm for AGV.Openmv camera as input,then arduino as controller also encoder for each mecanum wheels as feedback for positioning agv.Btw can i make that openmv determine the spool position either left or right side plus with distance from camera(already done) to tell AGV position(Camera) against spool(target) ?

[iabdalkader]

@aqeelyaacob Please post a new topic for that with as much details as possible.

[Mars]

Hi, I'm not sure why it's not working, try to remove the stages argument to test all stages. If you can you attach the xml and binary cascades I'll test them for you.

I use my own haarcascade_eye.cascade to detect eye
Error: unaligned read buf:10004fd6 count3


best wishes
Thankyou

[Falong]

Hi download the script like so:

Code: Select all

wget https://raw.githubusercontent.com/openmv/openmv/master/usr/openmv-cascade.py

The link above seems not existed any longer. I get it from https://github.com/openmv/openmv/blob/m ... convert.py
but when I ran it with attached xml file on ubuntu, it printed below error message:

Converting new XML format..
Traceback (most recent call last):
File "cascade_convert.py", line 422, in <module>
main()
File "cascade_convert.py", line 419, in main
cascade_binary_universal(args.file, args.stages, args.name)
File "cascade_convert.py", line 101, in cascade_binary_universal
cascade_binary(path, n_stages, name)
File "cascade_convert.py", line 183, in cascade_binary
rects = feature[idx].getElementsByTagName('_')
IndexError: list index out of range

[iabdalkader]

Hi download the script like so:

Code: Select all

wget https://raw.githubusercontent.com/openmv/openmv/master/usr/openmv-cascade.py

The link above seems not existed any longer. I get it from https://github.com/openmv/openmv/blob/m ... convert.py
but when I ran it with attached xml file on ubuntu, it printed below error message:

Converting new XML format..
Traceback (most recent call last):
File "cascade_convert.py", line 422, in <module>
main()
File "cascade_convert.py", line 419, in main
cascade_binary_universal(args.file, args.stages, args.name)
File "cascade_convert.py", line 101, in cascade_binary_universal
cascade_binary(path, n_stages, name)
File "cascade_convert.py", line 183, in cascade_binary
rects = feature[idx].getElementsByTagName('_')
IndexError: list index out of range

Hi, support for the new Haar cascade format is somewhat limited, we only support 1 feature and 2 alpha per weak classifier. This cascade has 2 features and 3 alpha per weak classifier so it's not going to work. See the haarcascade_frontalcatface.xml for an example of a supported new format cascade, or use the old format.

[latheefth]

Hi,
Does the cascade converter work on macos?I couldnt find the eye cascade binary in the Github directory, so had to generate it but it doesnt detect. I used the frontalface cascade binary provided and it works.

[kwagyeman]

Hi, I'll put some effort into updating the docs for all that in the future. Um, but, yest, it works on Mac if you have all the required libs.

[latheefth]

Thanks for the reply. If the binary generated from the provided eye cascade xml doesn't detect the eye, what could be the problem? i pass roi from the frontalface detect. I have attached my files. Could you please share me a working eye cascade binary, i am confused?

[iabdalkader]

Does the cascade converter work on macos?

Haven't tested it before on Mac.

I couldnt find the eye cascade binary in the Github directory

The eye and face cascades are built-in the firmware because they are very common. To load a built-in cascade, just use its name without a path, ex:

Code: Select all

c = load('eye.cascade')

. Also, you can re-generate any cascade from xml using the script.

If the binary generated from the provided eye cascade xml doesn't detect the eye, what could be the problem? i pass roi from the frontalface detect. I have attached my files. Could you please share me a working eye cascade binary, i am confused?

Could you try the example first using the built-in cascades, and see if it works ?

Examples->08-Eye-Tracking->face_eye_detection.py

[latheefth]

Hi,
face_eye_detection.py looks fine but it doesnt work with the eye cascade generated by me. I have attached the cascade in my previous post, could you please regenerate and try?
Thanks for the support. Much appreciated!

[iabdalkader]

Hi,
face_eye_detection.py looks fine but it doesnt work with the eye cascade generated by me. I have attached the cascade in my previous post, could you please regenerate and try?
Thanks for the support. Much appreciated!

The script should generate an identical cascade. Note I never tested the converted on Mac (and it doesn't work on Windows) I can only confirm it works on Linux, you could try to run the script on Linux VM and compare the output.

[augur4u]

Hi,
face_eye_detection.py looks fine but it doesnt work with the eye cascade generated by me. I have attached the cascade in my previous post, could you please regenerate and try?
Thanks for the support. Much appreciated!

The script should generate an identical cascade. Note I never tested the converted on Mac (and it doesn't work on Windows) I can only confirm it works on Linux, you could try to run the script on Linux VM and compare the output.

Hi iabdalkader,

I tried to use your cascade_convert.py to convert haarcascade_frontalcatface.xml but always failed. and I don't know why.
could you help me convert it?
Thanks a ton!