-
Notifications
You must be signed in to change notification settings - Fork 2
/
app_pp_reader.py
111 lines (83 loc) · 4.71 KB
/
app_pp_reader.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
import time
import cv2
import numpy as np
import logging
from kernel.interface.pp_reader import GetHandsInfo, get_fps_text
from kernel.media.video_processor import get_video_stream, get_mp4_video_writer, frame_operation
class PPReaderDemo:
def __init__(self, video_path, device, window_w=960, window_h=720, out_fps=18):
self.window_w = window_w
self.window_h = window_h
self.out_fps = out_fps
self.video_cap = get_video_stream(video_path)
self.pp_reader = GetHandsInfo(device, window_w, window_h)
self.image = None
def frame_processor(self):
if self.pp_reader.results is None:
logging.info("process result is null, please call hands_model_process function firstly!")
if self.pp_reader.results.multi_hand_landmarks:
handedness_list = self.pp_reader.check_hands_index(self.pp_reader.results.multi_handedness)
hand_num = len(handedness_list)
self.pp_reader.mode_processor.hand_num = hand_num
frame_copy = self.image.copy()
for hand_index, hand_landmarks in enumerate(self.pp_reader.results.multi_hand_landmarks):
if hand_index > 1:
hand_index = 1
# Label fingers
self.pp_reader.draw_hands_mark(self.image, hand_landmarks)
index_finger_tip_x, index_finger_tip_y = self.pp_reader.get_index_finger_tip_axis(hand_landmarks.landmark)
self.image = self.pp_reader.draw_paw_box(self.image, hand_landmarks.landmark, handedness_list, hand_index)
self.image = self.pp_reader.mode_processor.mode_execute(handedness_list[hand_index],
[index_finger_tip_x, index_finger_tip_y],
self.image, frame_copy)
else:
self.pp_reader.mode_processor.none_mode()
return self.image
def generate_pp_reader(self):
# using time to calculate fps
fps_time = time.time()
video_writer = get_mp4_video_writer(self.out_fps, self.window_w, self.window_h)
while self.video_cap.isOpened():
success, self.image = self.video_cap.read()
if not success:
print("空帧.")
continue
self.image = cv2.resize(self.image, (self.window_w, self.window_h))
# close write mode to improve performance
self.image.flags.writeable = False
# rotate or flip or None
self.image = frame_operation(self.image, rotate=False, flip=False)
# mediapipe mode process
self.pp_reader.hands_model_process(self.image)
self.image.flags.writeable = True
self.image = cv2.cvtColor(self.image, cv2.COLOR_RGB2BGR)
# read
# self.pp_reader.mode_processor.reader()
# Todo: store thumbnail, this function need to realize this file
if isinstance(self.pp_reader.mode_processor.last_thumb_img, np.ndarray):
self.image = self.pp_reader.mode_processor.generate_thumbnail(
self.pp_reader.mode_processor.last_thumb_img, self.image)
self.image = self.frame_processor()
# 显示刷新率FPS
ctime = time.time()
fps_text = get_fps_text(ctime, fps_time)
fps_time = ctime
self.image = self.pp_reader.mode_processor.generator.add_text(self.image, "帧率: " + str(int(fps_text)),
(10, 30), text_color=(0, 255, 0), text_size=50)
self.image = self.pp_reader.mode_processor.generator.add_text(self.image, "手掌: " +
str(self.pp_reader.mode_processor.hand_num),
(10, 90), text_color=(0, 255, 0), text_size=50)
self.image = self.pp_reader.mode_processor.generator.add_text(self.image, "模式: " +
str(self.pp_reader.mode_processor.hand_mode),
(10, 150), text_color=(0, 255, 0), text_size=50)
cv2.namedWindow('PPReader', cv2.WINDOW_FREERATIO)
cv2.imshow('PPReader', self.image)
video_writer.write(self.image)
# read
self.pp_reader.mode_processor.reader()
if cv2.waitKey(5) & 0xFF == 27:
break
self.video_cap.release()
if __name__ == '__main__':
pp_reader = PPReaderDemo(0, "CPU")
pp_reader.generate_pp_reader()