辨識標的物

[yillkid]

描述

根據原始碼：

• 連續 20~30 frames 都是某 label 的話，辨識確定
• 提示

frame_counts = 0
while True:
    frame_counts = frame_counts + 1

預期目標

• 辨識 label 內的標的物

[yillkid]

key_detect = 0
  times=1
  old_labels = ""
  frame_count = 0

  while (key_detect==0):
    ret,image_src =cap.read()

    image=cv2.resize(image_src,(224,224))

    start_time = time.time()
    results = classify_image(interpreter, image)
    elapsed_ms = (time.time() - start_time) * 1000
    label_id, prob = results[0]

    if frame_count==0:
        old_labels = labels[label_id]
        #print("Reset old_label")

    if old_labels == labels[label_id]:
        frame_count = frame_count+1
        #print("frame count ++")

    if frame_count>=30:
        print("Update old_label " + labels[label_id] + " and reset frame count")
        frame_count = 0

    print("frame_count:"+str(frame_count) +" tensor id:" + labels[label_id] + " and old id: " + str(old_labels))

    # cv2.imshow('Detecting....',image_src)

    if cv2.waitKey(1) & 0xFF == ord('q'):
      key_detect = 1

  cap.release()
  cv2.destroyAllWindows()

if __name__ == '__main__':
  main()

[yillkid]

參考程式碼：

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import argparse
import io
import time
import numpy as np
#import picamera
import cv2

from PIL import Image
from tflite_runtime.interpreter import Interpreter

def load_labels(path):
  with open(path, 'r') as f:
    return {i: line.strip() for i, line in enumerate(f.readlines())}

def set_input_tensor(interpreter, image):
  tensor_index = interpreter.get_input_details()[0]['index']
  input_tensor = interpreter.tensor(tensor_index)()[0]
  input_tensor[:, :] = image

def classify_image(interpreter, image, top_k=1):
  """Returns a sorted array of classification results."""
  set_input_tensor(interpreter, image)
  interpreter.invoke()
  output_details = interpreter.get_output_details()[0]
  output = np.squeeze(interpreter.get_tensor(output_details['index']))

  # If the model is quantized (uint8 data), then dequantize the results
  if output_details['dtype'] == np.uint8:
    scale, zero_point = output_details['quantization']
    output = scale * (output - zero_point)

  ordered = np.argpartition(-output, top_k)
  return [(i, output[i]) for i in ordered[:top_k]]

def main():
  parser = argparse.ArgumentParser(
      formatter_class=argparse.ArgumentDefaultsHelpFormatter)
  parser.add_argument(
      '--model', help='File path of .tflite file.', required=True)
  parser.add_argument(
      '--labels', help='File path of labels file.', required=True)
  args = parser.parse_args()

  labels = load_labels(args.labels)

  interpreter = Interpreter(args.model)
  interpreter.allocate_tensors()
  _, height, width, _ = interpreter.get_input_details()[0]['shape']

  cap = cv2.VideoCapture(0)
  # 擷取畫面 寬度 設定為512
  cap.set(cv2.CAP_PROP_FRAME_WIDTH,640)
  # 擷取畫面 高度 設定為512
  cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 480)

  key_detect = 0
  frame_count = 0
  diff_frame_count = 0

  defalut_labels = ""

  while (key_detect==0):
    ret,image_src =cap.read()

    image=cv2.resize(image_src,(224,224))

    start_time = time.time()
    results = classify_image(interpreter, image)
    elapsed_ms = (time.time() - start_time) * 1000
    label_id, prob = results[0]

    # 在 frame count 0 中定義 default label
    if frame_count == 0:
        default_labels = labels[label_id]

    # 如果 tensor 抓到的 label != default label, 開始計次
    if labels[label_id] != default_labels:
        print("Hello got different, diff_frame_count : " + str(diff_frame_count))
        diff_frame_count = diff_frame_count + 1
    elif labels[label_id] == default_labels:
        diff_frame_count = 0

    # 計次 30 次後，正式取代 default label
    if diff_frame_count == 10:
        print("Hello change default_label from " + default_labels + " to " + labels[label_id])
        default_labels = labels[label_id]
        diff_frame_count = 0

    time.sleep(3)
    frame_count = frame_count + 1

    print("Hello tensorflow got label : " + str(labels[label_id]))
    print("Hello default label : " + default_labels)

    if cv2.waitKey(1) & 0xFF == ord('q'):
      key_detect = 1

  cap.release()
  cv2.destroyAllWindows()

if __name__ == '__main__':
  main()

[yillkid]

karina :

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import argparse
import io
import time
import numpy as np
#import picamera
import cv2

from PIL import Image
from tflite_runtime.interpreter import Interpreter

def load_labels(path):
  with open(path, 'r') as f:
    return {i: line.strip() for i, line in enumerate(f.readlines())}

def set_input_tensor(interpreter, image):
  tensor_index = interpreter.get_input_details()[0]['index']
  input_tensor = interpreter.tensor(tensor_index)()[0]
  input_tensor[:, :] = image

def classify_image(interpreter, image, top_k=1):
  """Returns a sorted array of classification results."""
  set_input_tensor(interpreter, image)
  interpreter.invoke()
  output_details = interpreter.get_output_details()[0]
  output = np.squeeze(interpreter.get_tensor(output_details['index']))

  # If the model is quantized (uint8 data), then dequantize the results
  if output_details['dtype'] == np.uint8:
    scale, zero_point = output_details['quantization']
    output = scale * (output - zero_point)

  ordered = np.argpartition(-output, top_k)
  return [(i, output[i]) for i in ordered[:top_k]]

def main():
  parser = argparse.ArgumentParser(
      formatter_class=argparse.ArgumentDefaultsHelpFormatter)
  parser.add_argument(
      '--model', help='File path of .tflite file.', required=True)
  parser.add_argument(
      '--labels', help='File path of labels file.', required=True)
  args = parser.parse_args()

  labels = load_labels(args.labels)

  interpreter = Interpreter(args.model)
  interpreter.allocate_tensors()
  _, height, width, _ = interpreter.get_input_details()[0]['shape']

  #with picamera.PiCamera(resolution=(640, 480), framerate=30) as camera:
    #camera.start_preview()

  cap = cv2.VideoCapture(0)
  #擷取畫面 寬度 設定為512
  cap.set(cv2.CAP_PROP_FRAME_WIDTH,640)
  #擷取畫面 高度 設定為512

  cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 480)

  key_detect = 0
  times=1
  old_labels = ""
  frame_count = 0

  while (key_detect==0):
    ret,image_src =cap.read()

    image=cv2.resize(image_src,(224,224))

    start_time = time.time()
    results = classify_image(interpreter, image)
    elapsed_ms = (time.time() - start_time) * 1000
    label_id, prob = results[0]

    if frame_count==0:
        old_labels = labels[label_id]
        #print("Reset old_label")

    if old_labels == labels[label_id]:
        frame_count = frame_count+1
        #print("frame count ++")
    else:
        frame_count = 0

    if frame_count>=10:
        print("Update old_label " + labels[label_id] + " and reset frame count")
        frame_count = 0

    print("frame_count:"+str(frame_count) +" tensor id:" + labels[label_id] + " and old id: " + str(old_labels))

    # cv2.imshow('Detecting....',image_src)

    if cv2.waitKey(1) & 0xFF == ord('q'):
      key_detect = 1

  cap.release()
  cv2.destroyAllWindows()

if __name__ == '__main__':
  main()

[yillkid]

Yillkid 流程圖

karina1022 流程圖與討論

[yillkid]

已經提交程式碼。