28 Unstar Star 121 Fork 24

cungudafa / hand-keras-yolo3-recognize

Create your Gitee Account
Explore and code with more than 5 million developers,Free private repositories !:)
Sign up
Clone or download
pose_hand.py 10.52 KB
Copy Edit Web IDE Raw Blame History
cungudafa authored 2020-08-08 09:50 . sign
import time
import numpy as np
from pose.coco import general_coco_model
from pose.hand import general_hand_model
from pose.data_process import getBoneInformation, getHandsInformation,getPoseAndYoloInfo
from pose.hand_fD import hand_fourierDesciptor
from yolo import YOLO
from cv2 import cv2
from PIL import Image
import matplotlib.pyplot as plt
plt.rcParams['font.sans-serif'] = ['SimHei'] # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False # 用来正常显示负号
def getImageInfo(img_file, modelpath):
"""获得所有信息
:param 图片,模型路径
:return list单图的信息
"""
print("[INFO]Pose estimation.")
start = time.time()
pose_model = general_coco_model(modelpath) # 1.加载模型
print("[INFO]Model loads time: ", time.time() - start)
# 骨骼
start = time.time()
img = cv2.imread(img_file)
bone_points = pose_model.getBoneKeypoints(img) # 2.骨骼关键点
print("[INFO]COCO18_Model predicts time: ", time.time() - start)
lineimage,dotimage,black_np = pose_model.vis_bone_pose(img, bone_points) # 骨骼连线图、标记图显示cv2格式
list1 = getBoneInformation(bone_points) # 3.骨骼特征
#print("[INFO]Model Bone Information[25]: ", list1)
# yolo
_yolo = YOLO()
# cv2图片转PIL
image = Image.open(img_file)
lineimage = Image.fromarray(cv2.cvtColor(lineimage,cv2.COLOR_BGR2RGB))
black_np = Image.fromarray(cv2.cvtColor(black_np,cv2.COLOR_BGR2RGB))
r_image,labelinfo,hand_ROI_PIL = _yolo.detect_image(image,black_np) # 原图,lineimage线图,黑幕图
# plt.figure(figsize=[5, 5])
# plt.subplot(1, 2, 1)
# plt.imshow(r_image)
# plt.xlabel(u'线图', fontsize=20)
# plt.axis("off")
# plt.subplot(1, 2, 2)
# plt.imshow(cv2.cvtColor(dotimage, cv2.COLOR_BGR2RGB))
# plt.xlabel(u'点图', fontsize=20)
# plt.axis("off")
# plt.show()
# # 手势
# print("[INFO]Hands estimation.by handpose")
# start = time.time()
# hand_model = general_hand_model(modelpath) # 1.加载模型
# hand_fd = hand_fourierDesciptor()
# for i,handimg in enumerate(hand_ROI_PIL):
# img = cv2.cvtColor(np.asarray(handimg),cv2.COLOR_RGB2BGR)
# # hand 模型
# onehandpoints = hand_model.getOneHandKeypoints(img)
# hand_model.vis_hand_pose(img, onehandpoints)# 显示
# # hand_FD 描述子
# res1 = hand_fd.skinMask(img) # 进行肤色检测
# ret1, fourier = hand_fd.fourierDesciptor(res1) # 傅里叶描述子获取轮廓点
info = []
for i in range(len(list1)):
info.append(list1[i])
for j in range(len(labelinfo)):
info.append(labelinfo[j])
print(labelinfo)
return r_image,info
def getVideo_Info(modelpath,video_path, output_path=""):
from cv2 import cv2
cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
raise IOError("Couldn't open webcam or video")
video_FourCC = int(cap.get(cv2.CAP_PROP_FOURCC)) # 获取原始视频的信息
video_fps = cap.get(cv2.CAP_PROP_FPS)
video_size = (int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)),
int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)))
isOutput = True if output_path != "" else False # 如果设置了视频保存路径,则保存视频
if isOutput:
print("!!! TYPE:", type(output_path), type(
video_FourCC), type(video_fps), type(video_size))
out = cv2.VideoWriter(output_path, video_FourCC,
video_fps, video_size) # 根据原视频设置 保存视频的路径、大小、帧数
accum_time = 0
curr_fps = 0
fps = "FPS: ??"
prev_time = time.time()
yolo = YOLO()
pose_model = general_coco_model(modelpath) # 1.加载模型
while True:
return_value, frame = cap.read()
if return_value:
# 骨骼
bone_points = pose_model.getBoneKeypoints(frame) # 2.骨骼关键点
lineimage,dotimage = pose_model.vis_bone_pose(frame, bone_points) # 骨骼连线图、标记图显示
# list1 = getBoneInformation(bone_points) # 3.骨骼特征
temp,labelinfo = yolo.detect_image(Image.fromarray(frame),Image.fromarray(lineimage)) # 检测PIL格式
result = np.asarray(temp) # 画图到全部图上
curr_time = time.time()
exec_time = curr_time - prev_time
prev_time = curr_time
accum_time = accum_time + exec_time
curr_fps = curr_fps + 1
if accum_time > 1:
accum_time = accum_time - 1
fps = "FPS: " + str(curr_fps)
curr_fps = 0
cv2.putText(result, text=fps, org=(3, 15), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.50, color=(255, 0, 0), thickness=2)
cv2.putText(result, "q-'quit'", org=(3, 45), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.50, color=(0, 255, 0), thickness=2) # 标注字体
cv2.namedWindow("result", cv2.WINDOW_NORMAL)
cv2.imshow("result", result)
if isOutput:
out.write(result)
else:
print("Frame is end!")
break
if cv2.waitKey(1) & 0xFF == ord('q'):
break
def getImgInfo(img,pose_model,_yolo,save_path=""):
# 读取图像
#image = cv2.imread(img_path)
# 图像像素大小一致,视频保存尺寸有所冲突,这里弃用
#img = cv2.resize(img, (256, 256), interpolation=cv2.INTER_CUBIC)
""" ################关键点坐标检测##################"""
# pose骨骼
start = time.time()
bone_points = pose_model.getBoneKeypoints(img) # 1.骨骼关键点
#list1 = getBoneInformation(bone_points) # 骨骼特征
yololabel = _yolo.getyoloPoints(Image.fromarray(img)) # 2.yolo关键点
print("[INFO]Model predicts time: ", time.time() - start)
# print(bone_points)
# print(yololabel)
""" ################关键点距离角度特征信息##################"""
list1=0
list1 = getPoseAndYoloInfo(bone_points,yololabel) # 全部特征
# print(list1)
""" ################绘图##################"""
lineimage,dotimage,black_np = pose_model.vis_bone_pose(img, bone_points) # pose绘图
lineimage,dotimage,black_np = _yolo.vis_hand_pose(lineimage, dotimage, black_np, bone_points, yololabel) # yolo绘图
# lineimage = Image.fromarray(cv2.cvtColor(lineimage,cv2.COLOR_BGR2RGB))# yolo绘图
# black_np = Image.fromarray(cv2.cvtColor(black_np,cv2.COLOR_BGR2RGB))
# line_image,yololabel,hand_ROI_PIL = _yolo.detect_image(image,black_np)
isSave = True if save_path != "" else False # 如果设置了视频保存路径,则保存视频
if isSave:
#lineimage.save(save_path)
cv2.imwrite(save_path,lineimage,[int(cv2.IMWRITE_JPEG_QUALITY),70])
#cv2.imwrite('1.png', img, [int(cv2.IMWRITE_JPEG_QUALITY),95])#图像的质量,用0 - 100的整数表示,默认95;对于png ,第三个参数表示的是压缩级别。默认为3.
#cv2.imwrite('1.png',img,[int(cv2.IMWRITE_PNG_COMPRESSION),9])#从0到9 压缩级别越高图像越小
# plt.figure(figsize=[5, 5])
# plt.subplot(1, 3, 1)
# plt.imshow(cv2.cvtColor(lineimage, cv2.COLOR_BGR2RGB))
# plt.xlabel(u'线图', fontsize=20)
# plt.axis("off")
# plt.subplot(1, 3, 2)
# plt.imshow(cv2.cvtColor(dotimage, cv2.COLOR_BGR2RGB))
# plt.xlabel(u'点图', fontsize=20)
# plt.axis("off")
# plt.subplot(1, 3, 3)
# plt.imshow(cv2.cvtColor(black_np, cv2.COLOR_BGR2RGB))
# plt.xlabel(u'关键点线图', fontsize=20)
# plt.axis("off")
# plt.show()
return list1, lineimage
def getVideoInfo(video_path,pose_model,yolo,output_path=""):
from cv2 import cv2
cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
raise IOError("Couldn't open webcam or video")
fourcc = int(cap.get(cv2.CAP_PROP_FOURCC)) # 获取原始视频的信息
fps = cap.get(cv2.CAP_PROP_FPS)
size = (int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)),int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)))
isOutput = True if output_path != "" else False # 如果设置了视频保存路径,则保存视频
if isOutput:
print("[INFO] video TYPE:", type(output_path), type(fourcc), type(fps), type(size))
out = cv2.VideoWriter(output_path, fourcc, fps, size) # 根据原视频设置 保存视频的路径、大小、帧数
accum_time = 0
curr_fps = 0
fps = "FPS: ??"
prev_time = time.time()
while True:
ret, frame = cap.read()
if ret:
info, lineimage = getImgInfo(frame, pose_model, yolo) # 每一帧检测
curr_time = time.time()
exec_time = curr_time - prev_time
prev_time = curr_time
accum_time = accum_time + exec_time
curr_fps = curr_fps + 1
if accum_time > 1:
accum_time = accum_time - 1
fps = "FPS: " + str(curr_fps)
curr_fps = 0
cv2.putText(lineimage, text=fps, org=(3, 15), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.50, color=(255, 0, 0), thickness=2)
cv2.putText(lineimage, "q-'quit'", org=(3, 45), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.50, color=(0, 255, 0), thickness=2) # 标注字体
cv2.namedWindow("result", cv2.WINDOW_NORMAL)
cv2.imshow("result", lineimage)
if isOutput:
out.write(lineimage)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
out.release()
cv2.destroyAllWindows()
if __name__ == '__main__':
# coco
modelpath = "model/"
start = time.time()
pose_model = general_coco_model(modelpath) # 1.加载模型
print("[INFO]Pose Model loads time: ", time.time() - start)
# yolo
start = time.time()
_yolo = YOLO() # 1.加载模型
print("[INFO]yolo Model loads time: ", time.time() - start)
img_path = 'docs/classmates_5_2.png'
img_path = 'docs/wangyu_hand_img/brave_92.jpg'
image = cv2.imread(img_path)
#getImgInfo(image,pose_model,_yolo,'docs/brave_92_lineimg.jpg')
#getVideoInfo("docs/sun.mp4",pose_model,_yolo,"docs/sun_detect.mp4")
getVideoInfo("D:/myworkspace/dataset/My_test/video/wy2/you.mp4",pose_model,_yolo,"D:/myworkspace/dataset/My_test/you_detect.mp4")

Comment ( 0 )

Sign in for post a comment

Python
1
https://gitee.com/cungudafa/hand-keras-yolo3-recognize.git
git@gitee.com:cungudafa/hand-keras-yolo3-recognize.git
cungudafa
hand-keras-yolo3-recognize
hand-keras-yolo3-recognize
master

Search

132457 8cb2edc1 1899542 131848 70c8d3a4 1899542