Worked yesterday, today ValueError: Could not open '~/google-coral/examples-camera/all_models/mobilenet_ssd_v1_face_quant_postprocess_edgetpu.tflite'.

[lamachine]

Description

I was running the last several days modifying opencv/detect.py for three usb cameras. All went well. I also modified it to just gather images with no inference. Grabbed a couple thousand images yesterday and tried to post-process and got the above error. Today same thing. I deleted and reinstalled the entire camera examples folder including downloading all models again. No joy.

Changed models to v1 and to face, same issue.

RPi4, coral usb plugged into blue jacks, updated raspbian etc (like I said, worked yesterday).

Since it worked yesterday, I don't think it is a bug in this code, but I am at a loss of what it could be. I am running it from a separate folder but changed the model folder to the correct one. See below.

This works:
`
import cv2
import time
import datetime

import argparse
import os

def main():
default_save_dir = './images'

#for i in range (1,6):
while True:
    cameras = (0, 2, 4)
    for camera in cameras:
        
        # Create a VideoCapture object
        cap = cv2.VideoCapture(camera)
        # Check if the camera is opened successfully
        if not cap.isOpened():
            print("Could not open camera")
            exit()
        # Capture a frame
        ret, frame = cap.read()
        file_name = ("image"+ str(camera) + str(datetime.datetime.now()) +".jpg")
        complete_name = os.path.join(default_save_dir, file_name)
        cv2.imwrite(complete_name, frame)
        
        
        cap.release()

if name == 'main':
main()
`

This does not. Error below
`

v4l2-ctl --list-devices

import cv2
import time
import datetime

import argparse

import os

from pycoral.adapters.common import input_size
from pycoral.adapters.detect import get_objects
from pycoral.utils.dataset import read_label_file
from pycoral.utils.edgetpu import make_interpreter
from pycoral.utils.edgetpu import run_inference

def main():
default_model_dir = '~/google-coral/examples-camera/all_models'
default_model = 'mobilenet_ssd_v1_face_quant_postprocess_edgetpu.tflite'
default_labels = 'coco_labels.txt'
parser = argparse.ArgumentParser()
parser.add_argument('--model', help='.tflite model path',
default=os.path.join(default_model_dir,default_model))
parser.add_argument('--labels', help='label file path',
default=os.path.join(default_model_dir, default_labels))
parser.add_argument('--top_k', type=int, default=3,
help='number of categories with highest score to display')
parser.add_argument('--camera_idx', type=int, help='Index of which video source to use. ', default = 0)
parser.add_argument('--threshold', type=float, default=0.1,
help='classifier score threshold')
args = parser.parse_args()
default_save_dir = './images'

print('Loading {} with {} labels.'.format(args.model, args.labels))
interpreter = make_interpreter(args.model)
interpreter.allocate_tensors()
labels = read_label_file(args.labels)
inference_size = input_size(interpreter)



#for i in range (1,6):
while True:
    cameras = (0, 2, 4)
    for camera in cameras:
        
        # Create a VideoCapture object
        cap = cv2.VideoCapture(camera)
        # Check if the camera is opened successfully
        if not cap.isOpened():
            print("Could not open camera")
            exit()
        # Capture a frame
        ret, frame = cap.read()
        file_name = ("image"+ str(camera) + str(datetime.datetime.now()) +".jpg")
        complete_name = os.path.join(default_save_dir, file_name)
        cv2.imwrite(complete_name, frame)
        

        cv2_im_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
        cv2_im_rgb = cv2.resize(cv2_im_rgb, inference_size)
        run_inference(interpreter, cv2_im_rgb.tobytes())
        objs = get_objects(interpreter, args.threshold)[:args.top_k]
        final = append_objs_to_img(frame, inference_size, objs, labels)
        file_name = ("final"+ str(camera) + str(datetime.datetime.now()) +".jpg")
        complete_name = os.path.join(default_save_dir, file_name)
        cv2.imwrite(complete_name, frame)
        
        cap.release()

def append_objs_to_img(cv2_im, inference_size, objs, labels):
height, width, channels = cv2_im.shape
scale_x, scale_y = width / inference_size[0], height / inference_size[1]
for obj in objs:
bbox = obj.bbox.scale(scale_x, scale_y)
x0, y0 = int(bbox.xmin), int(bbox.ymin)
x1, y1 = int(bbox.xmax), int(bbox.ymax)
`

    percent = int(100 * obj.score)
    label = '{}% {}'.format(percent, labels.get(obj.id, obj.id))

    cv2_im = cv2.rectangle(cv2_im, (x0, y0), (x1, y1), (0, 255, 0), 2)
    cv2_im = cv2.putText(cv2_im, label, (x0, y0+30),
                         cv2.FONT_HERSHEY_SIMPLEX, 1.0, (255, 0, 0), 2)
return cv2_im

if name == 'main':
main()

Click to expand!

Issue Type

Support

Operating System

Ubuntu

Coral Device

USB Accelerator

Other Devices

Rapsberry Pi 4

Programming Language

No response

Relevant Log Output

python3 image_capture_ai.py 
Loading ~/google-coral/examples-camera/all_models/mobilenet_ssd_v1_face_quant_postprocess_edgetpu.tflite with ~/google-coral/examples-camera/all_models/coco_labels.txt labels.
Traceback (most recent call last):
  File "image_capture_ai.py", line 95, in <module>
    main()
  File "image_capture_ai.py", line 40, in main
    interpreter = make_interpreter(args.model)
  File "/usr/lib/python3/dist-packages/pycoral/utils/edgetpu.py", line 93, in make_interpreter
    model_path=model_path_or_content, experimental_delegates=delegates)
  File "/usr/lib/python3/dist-packages/tflite_runtime/interpreter.py", line 351, in __init__
    experimental_preserve_all_tensors))
ValueError: Could not open '~/google-coral/examples-camera/all_models/mobilenet_ssd_v1_face_quant_postprocess_edgetpu.tflite'.
pi@AshSjDemoPi40:~/Desktop/multiple-cameras-rpi $ dir ~/google-coral/examples-camera/all_models/
coco_labels.txt
imagenet_labels.txt
inat_bird_labels.txt
inat_insect_labels.txt
inat_plant_labels.txt
inception_v1_224_quant_edgetpu.tflite
inception_v1_224_quant.tflite
inception_v2_224_quant_edgetpu.tflite
inception_v2_224_quant.tflite
inception_v3_299_quant_edgetpu.tflite
inception_v3_299_quant.tflite
inception_v4_299_quant_edgetpu.tflite
inception_v4_299_quant.tflite
mobilenet_ssd_v1_coco_quant_postprocess_edgetpu.tflite
mobilenet_ssd_v1_coco_quant_postprocess.tflite
mobilenet_ssd_v2_coco_quant_postprocess_edgetpu.tflite
mobilenet_ssd_v2_coco_quant_postprocess.tflite
mobilenet_ssd_v2_face_quant_postprocess_edgetpu.tflite
mobilenet_ssd_v2_face_quant_postprocess.tflite
mobilenet_v1_1.0_224_quant_edgetpu.tflite
mobilenet_v1_1.0_224_quant_embedding_extractor_edgetpu.tflite
mobilenet_v1_1.0_224_quant_embedding_extractor.tflite
mobilenet_v1_1.0_224_quant.tflite
mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite
mobilenet_v2_1.0_224_inat_bird_quant.tflite
mobilenet_v2_1.0_224_inat_insect_quant_edgetpu.tflite
mobilenet_v2_1.0_224_inat_insect_quant.tflite
mobilenet_v2_1.0_224_inat_plant_quant_edgetpu.tflite
mobilenet_v2_1.0_224_inat_plant_quant.tflite
mobilenet_v2_1.0_224_quant_edgetpu.tflite
mobilenet_v2_1.0_224_quant.tflite
pet_labels.txt

[lamachine]

Update, I cannot install requirements (specifically Open CV) from the script.

pi@AshSjDemoPi42:~/coral/examples-camera/opencv $ bash install_requirements.sh 
Raspbian Version: 11
Looking in indexes: https://pypi.org/simple, https://www.piwheels.org/simple
ERROR: Could not find a version that satisfies the requirement opencv-contrib-python==4.1.0.25
ERROR: No matching distribution found for opencv-contrib-python==4.1.0.25

[lamachine]

Further update, new AI Maker Kit image written to SD card. Followed instructions carefully

pi@AshSjDemoPi42:~/aiy-maker-kit $ python3 run_tests.py 
--- Checking display ---
Found a display.

--- Checking required files ---
Found the required files.

--- Testing camera ---
Press Q to quit
Camera started successfully.
Closing video in... 4
(process:4434): Gtk-WARNING **: 12:36:05.270: Locale not supported by C library.
	Using the fallback 'C' locale.
Closing video in... 0
Camera is working.

--- Testing USB Accelerator ---
Loading a model...
USB Accelerator is working.

Everything look good!

then

pi@AshSjDemoPi42:~/aiy-maker-kit/examples $ python3 detect_faces.py 
Press Q to quit
Camera started successfully.

(process:4550): Gtk-WARNING **: 12:40:29.270: Locale not supported by C library.
	Using the fallback 'C' locale.

Face recognition worked fine.

Now my program...

pi@AshSjDemoPi42:~/Desktop/Ash Project $ python3 ai_image_capture.py 
Loading ~/aiy-maker-kit/examples/models/ssd_mobilenet_v2_coco_quant_postprocess_edgetpu.tflite with ~/aiy-maker-kit/examples/models/coco_labels.txt labels.
Traceback (most recent call last):
  File "ai_image_capture.py", line 111, in <module>
    main()
  File "ai_image_capture.py", line 54, in main
    interpreter = make_interpreter(args.model)
  File "/usr/lib/python3/dist-packages/pycoral/utils/edgetpu.py", line 93, in make_interpreter
    model_path=model_path_or_content, experimental_delegates=delegates)
  File "/usr/lib/python3/dist-packages/tflite_runtime/interpreter.py", line 351, in __init__
    experimental_preserve_all_tensors))
ValueError: Could not open '~/aiy-maker-kit/examples/models/ssd_mobilenet_v2_coco_quant_postprocess_edgetpu.tflite'.

I thought maybe the path was bad, but checked with dir

pi@AshSjDemoPi42:~/Desktop/Ash Project $ dir ~/aiy-maker-kit/examples/models
coco_labels.txt
imagenet_labels.txt
mobilenet_v1_1.0_224_l2norm_quant_edgetpu.tflite
movenet_single_pose_lightning_ptq_edgetpu.tflite
ssd_mobilenet_v2_coco_quant_postprocess_edgetpu.tflite
ssd_mobilenet_v2_face_quant_postprocess_edgetpu.tflite
tf2_mobilenet_v2_1.0_224_ptq_edgetpu.tflite