def __init__(self, can, poids, coeff, py0, px0):
self.can = can
self.poids = poids
self.py0 = py0
self.px0 = px0
self.py = py0
self.px = px0
self.coeff = coeff
self.moov = 0
self.timer = MyTimer(0.001, self.deplace)
def get_camera_images(self):
'''
截取摄像头图片，存入全局变量中
'''
reactTimer = MyTimer()
waitTimer = MyTimer()
reactTime = 0.5
imageSize = (640, 480)
images = []
reactTime = int(reactTime * 1000)
reactTimer.start(reactTime)
waitTime = 100
while reactTimer.isTimeOut() == False:
waitTimer.start(waitTime)
flag, image = self.cap.read()
image = cv2.resize(image, imageSize)
images.append(image)
waitTimer.waitTimeOut()
gl.GL_Images = images
gl.GL_Images_Length = images.__len__()
print("length={}".format(images.__len__()))
lsqe = el.LSqEllipse()
lsqe.fit(data)
center, width, height, phi = lsqe.parameters()
a = max(width, height)
b = min(width, height)
area = math.pi * a * b
eccentricity = pow((1 - (b / a)**2), 1 / 2)
pore_opening_degree = b / a
return a, b, area, eccentricity, pore_opening_degree
if __name__ == "__main__":
import os
base_path = r"C:\Users\LJY\Desktop\第一篇小论文-第二阶段\cropped_stomata\cropped_det"
# 计时开始
mytime = MyTimer.MyTimer()
mytime.start()
# 先统计一下
path = r"D:\Mask_RCNN\mask_rcnn\cropped_stomata\first_time"
# image_num 是 f
image_num = len([
lists for lists in os.listdir(path)
if os.path.isdir(os.path.join(path, lists))
])
# pore_num 是 s，表示总共有多少个匹配气孔
pore_num = 0
for img in range(image_num):
dirname = r"D:\Mask_RCNN\mask_rcnn\cropped_stomata\first_time\{}".format(
img)
filenum = len([
print __name__, "start ..." max = 20 fig = plt.figure() ax = fig.add_subplot(111) # some X and Y data li, = ax.plot(x, y, 'r.') # draw and show it fig.canvas.draw() plt.ylim((-5,25)) plt.show(block=False) HOST, PORT = "192.168.0.123", 9998 timer = MyTimer.MyTimer() print timer print "start server" # Create the server, binding to localhost on port 9999 server = SocketServer.TCPServer((HOST, PORT), MyTCPHandler) server_thread = threading.Thread(target=server.serve_forever) # Exit the server thread when the main thread terminates server_thread.daemon = True server_thread.start() print "start drawer" drawer = PlotDrawer() drawer_thread = threading.Thread(target=drawer.run) # Exit the server thread when the main thread terminates
def run(self):
#获取全局变量数据并更新全局数据
self.info = gl.GL_VInfo
self.model = gl.GL_Model
self.isEnd = False
gl.GL_TalScore = 0
#清空全局队列
while gl.GL_CurScoreQueue.empty() == False:
gl.GL_CurScoreQueue.get()
startTimer = MyTimer()
startTimer.start(1000) #开始计时
temp = 0
for oneInfo in self.info:
timeid = oneInfo['timeid'] #演示视频播放时间ms
reactime = float(oneInfo['reactime']) #反应时间
poseSocre = oneInfo['score'] #分数
posepoints = oneInfo['posepoints'] #关节点信息
posepoints = mxnet.nd.array(posepoints) #类型转换
temp += 1
print('*************这是第{}次'.format(temp))
print("timeid={}".format(timeid))
startTimer.resetWaitTime(timeid) #设置等待时间
startTimer.waitTimeOut() #等待时间，相当于 sleep
self.now_to_getImages.emit() #发送截取视频图像的信号
#等待截取的图片存入全局变量
while bool(gl.GL_Images == []) | bool(
gl.GL_Images.__len__() < gl.GL_Images_Length):
if self.isEnd == True: #当需要线程关闭时，退出
break
time.sleep(0.02)
if self.isEnd == True: #当需要线程关闭时，退出
break
images = gl.GL_Images #获取全局变量 (截取的图片list)
#跟新全局变量数据
gl.GL_Images = []
gl.GL_Images_Length = 0
use_gpu = False
maxScore = 0
score = 0
#给每一个图片匹配示范骨架，获取最大分数
for image in images:
if self.isEnd == True:
break
playerPre, playerImage = net.detection(self.model, image,
use_gpu) #模型预测骨架
if playerPre != None:
playerDis, playerSkeleton = tools.normalization(
playerPre['pred_coords'][0]) #结果归一化
score = tools.matching(playerSkeleton,
posepoints,
use_gpu=use_gpu) #计算匹配结果
if maxScore < score:
maxScore = score
if self.isEnd == True:
break
print("mxScore={}".format(maxScore))
#当分数超过10，认定匹配成功
if maxScore > 10:
#将分数更新在全局变量中
gl.GL_CurScoreQueue.put(poseSocre)
gl.GL_TalScore += poseSocre
self.update_date.emit(poseSocre) #发送信号，传递动作得分
else:
self.update_date.emit(0) #发送信号，传递动作得分
startTimer.end()
# |-test
# |-|-test
# |-|-|-***.png
# |-|=annotations
# |-|-|-***.json
dataset_test.load_coco(r"D:\Mask_RCNN\mask_rcnn\测试银杏的参数数据集\test2017",
"test2017")
dataset_test.prepare()
# image_ids 是记录test_datasets中共有多少张图片
image_ids = dataset_test.image_ids
IoU_all_photos = np.zeros((1, len(image_ids)))
all_statistic = np.zeros((2, len(image_ids)))
count_statistic = np.zeros((2, len(image_ids)))
# 计时
t1 = MyTimer.MyTimer()
t1.start()
# 测试集
for f in range(len(image_ids)):
image_id = f
info = dataset_test.image_info[image_id]
print("image ID: {}.{} ({}) {}".format(
info["source"], info["id"], image_id,
dataset_test.image_reference(image_id)))
# 原demo在使用图像之前已经对图像进行了裁剪 具体代码在model的1224行
image, image_meta, gt_class_id, gt_bbox, gt_mask = modellib.load_image_gt(
dataset_test, config, image_id, use_mini_mask=False)
# image 是 3D图像 [image]是4-D列表 len([image]) = 1 [image]代表把整个图像作为一个batch处理
results = model.detect([image], verbose=1)
# 这里的model.detect
