def main():
# load net
net_file = join(realpath(dirname(__file__)), 'SiamRPNBIG.model')
net = SiamRPNBIG()
net.load_state_dict(torch.load(net_file))
net.eval().cuda()
# warm up
for i in range(10):
net.temple(
torch.autograd.Variable(torch.FloatTensor(1, 3, 127, 127)).cuda())
net(torch.autograd.Variable(torch.FloatTensor(1, 3, 255, 255)).cuda())
# start to track
handle = vot.VOT("polygon")
Polygon = handle.region()
cx, cy, w, h = get_axis_aligned_bbox(Polygon)
image_file = handle.frame()
if not image_file:
sys.exit(0)
target_pos, target_sz = np.array([cx, cy]), np.array([w, h])
im = cv2.imread(image_file) # HxWxC
state = SiamRPN_init(im, target_pos, target_sz, net) # init tracker
while True:
image_file = handle.frame()
if not image_file:
break
im = cv2.imread(image_file) # HxWxC
state = SiamRPN_track(state, im) # track
res = cxy_wh_2_rect(state['target_pos'], state['target_sz'])
handle.report(Rectangle(res[0], res[1], res[2], res[3]))
def track(self, image):
test_crop = utils.get_subwindow(image, self.pos, self.patch_size)
z = np.multiply(test_crop - test_crop.mean(), self.cos_window[:, :,
None])
zf = np.fft.fft2(z, axes=(0, 1))
k_test = utils.dense_gauss_kernel(self.feature_bandwidth_sigma,
self.xf, self.x, zf, z)
kf_test = np.fft.fft2(k_test, axes=(0, 1))
response = np.real(np.fft.ifft2(np.multiply(self.alphaf, kf_test)))
# Max position in response map
v_centre, h_centre = np.unravel_index(response.argmax(),
response.shape)
vert_delta, horiz_delta = [
v_centre - response.shape[0] / 2, h_centre - response.shape[1] / 2
]
# Predicted position
self.pos = [self.pos[0] + vert_delta, self.pos[1] + horiz_delta]
self.bbox = [
self.pos[1] - self.target_size[1] / 2,
self.pos[0] - self.target_size[0] / 2, self.target_size[1],
self.target_size[0]
]
self._results.append(self.bbox)
return Rectangle(self.pos[1] - self.target_size[1] / 2,
self.pos[0] - self.target_size[0] / 2,
self.target_size[1], self.target_size[0])
def track_vot2(self, imgtype):
"""Run tracker on a sequence."""
handle = vot.VOT("rectangle", "rgbt")
rect = list(handle.region())
colorimage, thermalimage = handle.frame()
startnum = 20
image_rgb = self._read_image(colorimage[startnum:len(colorimage)-2])
image_ir = self._read_image(thermalimage[startnum:len(thermalimage)-2])
if imgtype == 'rgb':
self.initialize(image_ir, image_rgb, rect)
elif imgtype == 'ir':
self.initialize(image_rgb, image_ir, rect)
while True:
colorimage, thermalimage = handle.frame()
if not colorimage:
break
image_rgb = self._read_image(colorimage[startnum:len(colorimage)-2])
image_ir = self._read_image(thermalimage[startnum:len(thermalimage)-2])
if imgtype == 'rgb':
state = self.track(image_ir, image_rgb)
elif imgtype == 'ir':
state = self.track(image_rgb, image_ir)
region = np.array(state).astype(int)
handle.report(Rectangle(region[0], region[1], region[2], region[3]))
net.eval().cuda()
# warm up
for i in range(10):
net.temple(torch.autograd.Variable(torch.FloatTensor(1, 3, 127, 127)).cuda())
net(torch.autograd.Variable(torch.FloatTensor(1, 3, 255, 255)).cuda())
# start to track
handle = vot.VOT("polygon")
Polygon = handle.region()
cx, cy, w, h = get_axis_aligned_bbox(Polygon)
image_file = handle.frame()
if not image_file:
sys.exit(0)
target_pos, target_sz = np.array([cx, cy]), np.array([w, h])
im = cv2.imread(image_file) # HxWxC
state = SiamRPN_init(im, target_pos, target_sz, net) # init tracker
while True:
image_file = handle.frame()
if not image_file:
break
im = cv2.imread(image_file) # HxWxC
state = SiamRPN_track(state, im) # track
# convert cx, cy, w, h into rect
res = cxy_wh_2_rect(state['target_pos'], state['target_sz'])
handle.report(Rectangle(res[0], res[1], res[2], res[3]))
SiamMask_dir = '/home/guo/zpy/Mypysot/mypysot/data_and_result/weight/Mob20.pth' #Res20.pth'
Dr_dir = '/home/guo/zpy/Mypysot/mypysot/data_and_result/weight/High-Low-two.pth.tar'
tracker = TSDMTracker(SiamRes_dir, SiamMask_dir, Dr_dir, image_rgb,
image_depth, region)
#track
while True:
image_file1, image_file2 = handle.frame()
if not image_file1:
break
image_rgb = cv2.imread(image_file1)
image_depth = cv2.imread(image_file2, -1)
state = tracker.track(image_rgb, image_depth)
region_Dr, score = state['region_Siam'], state[
'score'] #state['region_Dr']
if score > 0.3:
score = 1
else:
score = 0
handle.report(
Rectangle(region_Dr[0], region_Dr[1], region_Dr[2], region_Dr[3]),
score)
'''
handle = vot.VOT("rectangle",'rgbd')
region = handle.region()
image_file1, image_file2 = handle.frame()
while True:
handle.report(Rectangle(2,2,1,1), 1)
'''
except Exception as e:
logging.info('error %s', e)
while True:
c = c + 1
logging.info('************* %s *************', c)
image_file = handle.frame()
logging.info(image_file)
if not image_file:
break
# state = RTMDNet_track(state, image_file)
# box = state['target_bbox']
# #res = cxy_wh_2_rect(box)
# #handle.report(Rectangle(res[0], res[1], res[2], res[3]))
#handle.report(Rectangle(box[0], box[1], box[2], box[3]))
#logging.info('box: %s, res: %s', box, res)
# if c == 5:
# sys.exit(0)
try:
state = RTMDNet_track(state, image_file)
box = state['target_bbox']
#res = cxy_wh_2_rect(box)
#handle.report(Rectangle(res[0], res[1], res[2], res[3]))
handle.report(Rectangle(box[0], box[1], box[2], box[3]))
except Exception as e:
logging.info('error is: %s', e)
depth_im = np.array(Image.open(imagefile_d))
tracker.initialize(image=image, state=selection, depth_im=depth_im)
while True:
imagefile = handle.frame()
if not imagefile:
break
# imagefile = ''.join(imagefile)
# imagefile = imagefile[20:-2]
# image = tracker._read_image(imagefile)
imagefile_rgb = imagefile[0]
imagefile_d = imagefile[1]
assert imagefile_rgb.find("color") >= 0
assert imagefile_d.find("depth") >= 0
image = tracker._read_image(imagefile_rgb)
depth_im = np.array(Image.open(imagefile_d))
start_time = time.time()
state = tracker.track(image=image, depth_im=depth_im)
#overlay_boxes(image, state, imagefile[0], tracker_name)
#confidence=1
selection = Rectangle(state[0], state[1], state[2], state[3])
handle.report(selection)
time.sleep(0.01)
# for i in range(10):
# net.temple(torch.autograd.Variable(torch.FloatTensor(1, 3, 127, 127)).cuda()) # selonsy:cuda()表示使用GPU进行计算,FloatTensor(1, 3, 127, 127)：浮点型四维张量
# net(torch.autograd.Variable(torch.FloatTensor(1, 3, 255, 255)).cuda())
# start to track
handle = vot.VOT("polygon")
Polygon = handle.region() # region：将配置消息发送到客户端并接收初始化区域和第一个图像的路径。其返回值为初始化区域。
cx, cy, w, h = get_axis_aligned_bbox(Polygon) # get_axis_aligned_bbox：将坐标数据转换成 RPN 的格式
image_file = handle.frame() # frame 函数从客户端获取帧（图像路径）
if not image_file:
sys.exit(0)
target_pos, target_sz = np.array([cx, cy]), np.array([w, h])
im = cv2.imread(image_file) # HxWxC
state = SiamRPN_init(im, target_pos, target_sz, net) # init tracker,SiamRPN_init:构造状态结构体并运行模板分支
# 从第一帧开始跟踪，表示很奇怪，难道直接给定的不准确么？ # selonsy：改进点
while True: # 进入跟踪循环
image_file = handle.frame()
if not image_file:
break
im = cv2.imread(image_file) # HxWxC
state = SiamRPN_track(state, im) # track,SiamRPN_track:运行检测分支并更新状态变量
res = cxy_wh_2_rect(state['target_pos'], state['target_sz']) # cxy_wh_2_rect:将坐标转换成矩形框的表示形式
handle.report(Rectangle(res[0], res[1], res[2], res[3])) # report:将跟踪结果报告给客户端
print(handle.result)
print(handle.frames)
del handle