def __init__(self, arguments):
self.arguments = arguments
# Frame window dim
self.winWidth = 640
self.winHeight = 480
actionPredictor_params.__init__(self)
self.fps_time = 0
#self.step = 15
self.mode = {
'Pose Estimation': 'estimation',
'Tracking': 'tracking',
'Action Recognition': 'recognition'
}
w, h = model_wh(self.arguments.resize)
if w > 0 and h > 0:
self.estimator = TfPoseEstimator(get_graph_path(
self.arguments.model),
target_size=(w, h))
else:
self.estimator = TfPoseEstimator(get_graph_path(
self.arguments.model),
target_size=(432, 368))
self.cam = cv2.VideoCapture(self.arguments.camera)
# Tracker based on Sort
self.sort_max_age = 20
self.sort_min_hit = 3
self.tracker = Sort(self.sort_max_age, self.sort_min_hit)
def load_model():
global poseEstimator
global sk_cnn_actionPredicter
poseEstimator = TfPoseEstimator(get_graph_path('mobilenet_thin'),
target_size=(432, 368))
sk_cnn_actionPredicter = sk_cnn.SkelCNN()
sk_cnn_actionPredicter.load_weights(
'./action_pre_sk_cnn/models/skel_cnn_model/sk-cnn.hdf5')
def load_model():
global poseEstimator
poseEstimator = TfPoseEstimator(get_graph_path('mobilenet_thin'),
target_size=(432, 368))
# -*- coding: utf-8 -*-
"""
Created on Thu Oct 3 20:41:35 2019
@author: ASUS
"""
import cv2
import numpy as np
import settings
from pose.estimator import TfPoseEstimator
from pose.networks import get_graph_path
from imutils.video import VideoStream
poseEstimator = None
poseEstimator = TfPoseEstimator(get_graph_path('mobilenet_thin'), target_size=(432, 368))
cap=cv2.VideoCapture(0)
#cap = VideoStream(src='rtsp://*****:*****@192.168.51.162/PSIA/streaming/channels/102').start()
while True:
ret,frame=cap.read()
ret=True
if ret :
show = cv2.resize(frame, (settings.winWidth, settings.winHeight))
humans = poseEstimator.inference(show)
fps_time = 0
if __name__ == '__main__':
# 类别以及要保存的视频段长度
action = 'satnd'
clip_length = 90
root_path = '/home/dl1/datasets/actions/'
if not os.path.exists(root_path + action):
os.mkdir(root_path + action)
if not os.path.exists(root_path + action + '/txt/'):
os.mkdir(root_path + action + '/txt/')
os.mkdir(root_path + action + '/imgs/')
samples = len(os.listdir(root_path + action + '/txt/'))
sample_count = 1000 if samples == 0 else 1000 + samples
e = TfPoseEstimator(get_graph_path('mobilenet_thin'),
target_size=(432, 368))
cam = cv2.VideoCapture(0)
ret_val, image = cam.read()
joints = []
joints_imgs = []
while True:
ret_val, image = cam.read()
if ret_val:
humans = e.inference(image)
image, joint, *_, sk = TfPoseEstimator.get_humans(image,
humans,
imgcopy=False)
if joint:
if len(joints) < clip_length:
joints.append(joint[0])
