-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathDataMaker.py
215 lines (196 loc) · 11.6 KB
/
DataMaker.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
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
"""
批量制作训练数据集
输入图片路径
输出数据集文件
"""
import math
import os
import copy
from src import util
from src.body import Body
from output_processing import candidateright
import cv2
from PIL import Image
from Parents_demo.deeplab import DeeplabV3
from measure import Img
from output_processing import writedown_list
candidate_picture_file = r"images/dataset/action6/"
input_path = r"images/dataset/action6/"
save_path = "save_img"
excel_path = r"lengths.xlsx"
global r_image
if __name__ == "__main__":
body_estimation = Body('model/body_pose_model.pth') # 生成openpose模型
deeplab = DeeplabV3()
# hand_estimation = Hand('model/hand_pose_model.pth')
photopath = os.listdir(candidate_picture_file) # 绑定输入照片文件夹路径
num_picture = len(photopath)
###################################################################由此进入循环
excelpath = './'
npicture = 1
id_list=[]
tall_list=[]
Shoulder_list=[]
hip_list=[]
arm_list=[]
Forearm_list = []
thigh_list=[]
calf_list=[]
for i in range(num_picture):
test_image = photopath[i]
print(test_image)
oriImg = cv2.imread(candidate_picture_file + test_image) # B,G,R order
candidate, subset = body_estimation(oriImg) # 图片推理，
canvas = copy.deepcopy(oriImg)
canvas = util.draw_bodypose(canvas, candidate, subset)
# print(candidate)
# print(subset)
# print(np.shape(candidate))
candidateright(candidate, subset)
############################################################################
############################################################################
# -------------------------------------------------------------------------#
# 如果想要修改对应种类的颜色，到__init__函数里修改self.colors即可
# -------------------------------------------------------------------------#
# deeplab = DeeplabV3() #生成deeplab模型 与body_estimation一样不要重复使用########不在循环里
# ----------------------------------------------------------------------------------------------------------#
# mode用于指定测试的模式：
# 'predict' 表示单张图片预测，如果想对预测过程进行修改，如保存图片，截取对象等，可以先看下方详细的注释
# 'video' 表示视频检测，可调用摄像头或者视频进行检测，详情查看下方注释。
# 'fps' 表示测试fps，使用的图片是img里面的street.jpg，详情查看下方注释。
# 'dir_predict' 表示遍历文件夹进行检测并保存。默认遍历img文件夹，保存img_out文件夹，详情查看下方注释。
# 'export_onnx' 表示将模型导出为onnx，需要pytorch1.7.1以上。
# ----------------------------------------------------------------------------------------------------------#
mode = "predict"
# -------------------------------------------------------------------------#
# count 指定了是否进行目标的像素点计数（即面积）与比例计算
# name_classes 区分的种类，和json_to_dataset里面的一样，用于打印种类和数量
#
# count、name_classes仅在mode='predict'时有效
# -------------------------------------------------------------------------#
count = False
name_classes = ["background", "aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair",
"cow",
"diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train",
"tvmonitor"]
# name_classes = ["background","cat","dog"]
# ----------------------------------------------------------------------------------------------------------#
# video_path 用于指定视频的路径，当video_path=0时表示检测摄像头
# 想要检测视频，则设置如video_path = "xxx.mp4"即可，代表读取出根目录下的xxx.mp4文件。
# video_save_path 表示视频保存的路径，当video_save_path=""时表示不保存
# 想要保存视频，则设置如video_save_path = "yyy.mp4"即可，代表保存为根目录下的yyy.mp4文件。
# video_fps 用于保存的视频的fps
#
# video_path、video_save_path和video_fps仅在mode='video'时有效
# 保存视频时需要ctrl+c退出或者运行到最后一帧才会完成完整的保存步骤。
# ----------------------------------------------------------------------------------------------------------#
# ----------------------------------------------------------------------------------------------------------#
# test_interval 用于指定测量fps的时候，图片检测的次数。理论上test_interval越大，fps越准确。
# fps_image_path 用于指定测试的fps图片
#
# test_interval和fps_image_path仅在mode='fps'有效
# ----------------------------------------------------------------------------------------------------------#
test_interval = 100
fps_image_path = "img/street.jpg"
# -------------------------------------------------------------------------#
# dir_origin_path 指定了用于检测的图片的文件夹路径
# dir_save_path 指定了检测完图片的保存路径
#
# dir_origin_path和dir_save_path仅在mode='dir_predict'时有效
# -------------------------------------------------------------------------#
dir_origin_path = "img/"
dir_save_path = "img_out/"
# -------------------------------------------------------------------------#
# simplify 使用Simplify onnx
# onnx_save_path 指定了onnx的保存路径
# -------------------------------------------------------------------------#
simplify = True
onnx_save_path = "model_data/models.onnx"
if mode == "predict":
'''
predict.py有几个注意点
1、该代码无法直接进行批量预测，如果想要批量预测，可以利用os.listdir()遍历文件夹，利用Image.open打开图片文件进行预测。
具体流程可以参考get_miou_prediction.py，在get_miou_prediction.py即实现了遍历。
2、如果想要保存，利用r_image.save("img.jpg")即可保存。
3、如果想要原图和分割图不混合，可以把blend参数设置成False。
4、如果想根据mask获取对应的区域，可以参考detect_image函数中，利用预测结果绘图的部分，判断每一个像素点的种类，然后根据种类获取对应的部分。
seg_img = np.zeros((np.shape(pr)[0],np.shape(pr)[1],3))
for c in range(self.num_classes):
seg_img[:, :, 0] += ((pr == c)*( self.colors[c][0] )).astype('uint8')
seg_img[:, :, 1] += ((pr == c)*( self.colors[c][1] )).astype('uint8')
seg_img[:, :, 2] += ((pr == c)*( self.colors[c][2] )).astype('uint8')
'''
img = input_path + test_image # ---》
try:
image = Image.open(img)
except:
print('Open Error! Try again!')
else:
r_image = deeplab.detect_image(image, count=count, name_classes=name_classes) ################图片推理
r_image.save("img.jpg")
#r_image.show()
else:
raise AssertionError("Please specify the correct mode: 'predict', 'video', 'fps' or 'dir_predict'.")
img = Img("img.jpg")
# img.img_show()
p0, p1 = (int(candidate[0][0]), int(candidate[0][1])), (int(candidate[1][0]), int(candidate[1][1])) # 鼻子和脖子的连线
##--------------------------------------------------------以下为fusion monstedr 内容--------------------------------------------------------------------------##
p0, p1 = (int(candidate[0][0]), int(candidate[0][1])), (int(candidate[10][0]), int(candidate[10][1])) ###鼻子
p2, p5 = (int(candidate[2][0]), int(candidate[2][1])), (int(candidate[5][0]), int(candidate[5][1])) # 左右肩
p3, p6 = (int(candidate[3][0]), int(candidate[3][1])), (int(candidate[6][0]), int(candidate[6][1])) # 左右肘
p4, p7 = (int(candidate[4][0]), int(candidate[4][1])), (int(candidate[7][0]), int(candidate[7][1])) # 左右手腕
p8, p11 = (int(candidate[8][0]), int(candidate[8][1])), (int(candidate[11][0]), int(candidate[11][1])) ###腰围
p9, p12 = (int(candidate[9][0]), int(candidate[9][1])), (int(candidate[12][0]), int(candidate[12][1])) ### 左右膝盖
p10, p13 = (int(candidate[10][0]), int(candidate[10][1])), (
int(candidate[13][0]), int(candidate[13][1])) ###左右脚跟
# p1013=(int(int(candidate[10][0])+int(candidate[13][0]))/2,int(int(candidate[10][1])+int(candidate[13][1]))/2)
p1013 = (int((candidate[10][0] + candidate[13][0]) / 2), int((candidate[10][1] + candidate[13][1]) / 2))
print(p2, p5)
print(p8, p11)
print(p10, p13)
print(p1013)
# time_start = time.time()
tp0, tp1, length1 = img.measure(p0, p1013) ##身高
tp2, tp5, length2 = img.measure(p2, p5) # 肩宽 测量长度,输出边缘点
tp8, tp11, length3 = img.measure(p8, p11) ##臀围
length4 = int(
(math.sqrt(pow(candidate[2][0] - candidate[3][0], 2) + pow(candidate[2][1] - candidate[3][1], 2)) +
math.sqrt(pow(candidate[5][0] - candidate[6][0], 2) + pow(candidate[5][1] - candidate[6][1], 2))) / 2)
length5 = int(
(math.sqrt(pow(candidate[3][0] - candidate[4][0], 2) + pow(candidate[3][1] - candidate[4][1], 2)) +
math.sqrt(pow(candidate[6][0] - candidate[7][0], 2) + pow(candidate[6][1] - candidate[7][1], 2))) / 2)
length6 = int(
(math.sqrt(pow(candidate[8][0] - candidate[9][0], 2) + pow(candidate[8][1] - candidate[9][1], 2)) +
math.sqrt(pow(candidate[11][0] - candidate[12][0], 2) + pow(candidate[11][1] - candidate[12][1], 2))) / 2)
length7 = int(
(math.sqrt(pow(candidate[9][0] - candidate[10][0], 2) + pow(candidate[9][1] - candidate[10][1], 2)) +
math.sqrt(pow(candidate[12][0] - candidate[13][0], 2) + pow(candidate[12][1] - candidate[13][1], 2))) / 2)
print("身高", length1)
print("肩宽", length2)
print("臀宽", length3)
print("大臂", length4)
print("小臂", length5)
print("大腿", length6)
print("小腿", length7)
img.draw_point(p0)
img.draw_point(p1)
tp2 = img.measure_tall_one_point(p1)
img.draw_Crosspoint(tp2)
print("tp2" + str(tp2))
# img.img_show()
#tall_list.append(640-tp2[1])
tall_list.append(length1)
Shoulder_list.append(length2)
hip_list.append(length3)
arm_list.append(length4)
Forearm_list.append(length5)
thigh_list.append(length6)
calf_list.append(length7)
writedown_list(tall_list,'身高')
writedown_list(Shoulder_list, '肩宽')
writedown_list(hip_list, '臀宽')
writedown_list(arm_list, '大臂')
writedown_list(Forearm_list, '小臂')
writedown_list(thigh_list, '大腿')
writedown_list(calf_list, '小腿')
writedown_list(id_list,'id')