def generate_gt_anchor(img, box, anchor_width=16, draw_img_gt=None):
"""
将原先每张图的bbox标签转化为每个anchor标签
:param img: 输入的图像
:param box: bbox gt（8个数字表示一个框）
:param anchor_width:anchor的宽度
:param draw_img_gt:画出gt框后的图像
:return:
"""
if not isinstance(box[0], float):
box = [float(box[i]) for i in range(len(box))]
result = []
# 求解一个bbox下，能分解为多少个16宽度的小anchor，并求出最左和最右的小achor的id
left_anchor_num = int(math.floor(max(min(box[0], box[6]), 0) / anchor_width)) # float(x) 函数：对x向下取整
right_anchor_num = int(math.ceil(min(max(box[2], box[4]), img.shape[1]) / anchor_width)) # ceil(x) 函数返回一个大于或等于 x 的的最小整数
# 极端情况：右边的anchor超过了图像的宽度
if right_anchor_num * 16 + 15 > img.shape[1]:
right_anchor_num -= 1
# 将gt左右两边的anchor合并为一对
position_pair = [(i * anchor_width, (i + 1) * anchor_width - 1) for i in range(left_anchor_num, right_anchor_num)]
# 计算anchor的高度和中心是多少
y_top, y_bottom = cal_y_top_and_bottom(img, position_pair, box)
for i in range(len(position_pair)):
position = int(position_pair[i][0] / anchor_width) # anchor box的编号
h = y_bottom[i] - y_top[i] + 1 # anchor box的高度
cy = (float(y_bottom[i]) + float(y_top[i])) / 2.0 # anchor box的中心点
result.append((position, cy, h))
draw_img_gt = draw_image.draw_box_h_and_c(draw_img_gt, position, cy, h)
draw_img_gt = draw_image.draw_box_4pt(draw_img_gt, box, color=(0, 0, 255), thickness=1)
return result, draw_img_gt
def generate_gt_anchor(img, box, anchor_width=16, draw_img_gt=None):
"""
calsulate ground truth fine-scale box
:param img: input image
:param box: ground truth box (4 point)
:param anchor_width:
:return: tuple (position, h, cy)
"""
if not isinstance(box[0], float):
box = [float(box[i]) for i in range(len(box))]
result = []
# box数据格式为[x1 y1 x2 y2 x3 y3 x4 y4]，左上、左下、右上、右下
left_anchor_num = int(
math.floor(
max(min(box[0], box[6]), 0) /
anchor_width)) # the left side anchor of the text box, downwards
right_anchor_num = int(
math.ceil(
min(max(box[2], box[4]), img.shape[1]) /
anchor_width)) # the right side anchor of the text box, upwards
# handle extreme case, the right side anchor may exceed the image width
if right_anchor_num * 16 + 15 > img.shape[1]:
right_anchor_num -= 1
# combine the left-side and the right-side x_coordinate of a text anchor into one pair
# 把一个box中的每个anchor左右边界以元组形式保存在position_pair中
position_pair = [(i * anchor_width, (i + 1) * anchor_width - 1)
for i in range(left_anchor_num, right_anchor_num)]
# 计算每个gt anchor的真实位置，其实就是求解gt anchor的上边界和下边界
y_top, y_bottom = cal_y_top_and_bottom2(img, position_pair, box)
# y_top2, y_bottom2 = cal_y_top_and_bottom2(img, position_pair, box)
# for i in range(len(y_top)):
# if abs(y_top[i] - y_top2[i]) > 5 or abs(y_bottom[i] - y_bottom2[i]) > 5:
# print("**********")
# print(y_top[i],y_bottom[i],y_top2[i],y_bottom2[i])
# print()
#print("image shape: %s, pair_num: %s, top_num:%s, bot_num:%s" % (img.shape, len(position_pair), len(y_top), len(y_bottom)))
# 最后将每个anchor的位置(水平ID，从左到右第几个anchor)、anchor中心y坐标、anchor高度存储并返回
for i in range(len(position_pair)):
position = int(position_pair[i][0] /
anchor_width) # the index of anchor box
h = abs(y_bottom[i] - y_top[i]) + 1 # the height of anchor box
cy = (float(y_bottom[i]) +
float(y_top[i])) / 2.0 # the center point of anchor box
result.append(
(position, cy, h)) # result保存图片中box的所有ground truth anchor的水平位置、y、h
draw_img_gt = draw_image.draw_box_h_and_c(draw_img_gt, position, cy,
h) # 把anchor画出来
draw_img_gt = draw_image.draw_box_4pt(draw_img_gt,
box,
color=(0, 0, 255),
thickness=1)
# cv2.imwrite('a.jpg',draw_img_gt)
return result, draw_img_gt
def cal_y_top_and_bottom(raw_img, position_pair, box):
"""
:param raw_img:
:param position_pair: for example:[(0, 15), (16, 31), ...]
:param box: gt box (4 point)
:return: top and bottom coordinates for y-axis
"""
img = copy.deepcopy(raw_img)
y_top = []
y_bottom = []
height = img.shape[0]
img[:, :, 0] = 0
# for i in range(img.shape[0]):
# for j in range(img.shape[1]):
# img[i, j, 0] = 0
top_flag = False
bottom_flag = False
img = draw_image.draw_box_4pt(img, box, color=(255, 0, 0))
# calc top y coordinate, pixel from top to down loop
for k in range(len(position_pair)):
# calc top y coordinate
for y in range(0, height-1):
# loop each anchor, from left to right
for x in range(position_pair[k][0], position_pair[k][1] + 1):
if img[y, x, 0] == 255:
y_top.append(y)
top_flag = True
break
if top_flag is True:
break
# calc bottom y coordinate, pixel from down to top loop
for y in range(height - 1, -1, -1):
# loop each anchor, from left to right
for x in range(position_pair[k][0], position_pair[k][1] + 1):
if img[y, x, 0] == 255:
y_bottom.append(y)
bottom_flag = True
break
if bottom_flag is True:
break
top_flag = False
bottom_flag = False
# print(y_top)
# print(y_bottom)
return y_top, y_bottom
def generate_gt_anchor(img, box, anchor_width=16, draw_img_gt=None):
"""
calsulate ground truth fine-scale box
:param img: input image
:param box: ground truth box (4 point)
:param anchor_width:
:return: tuple (position, h, cy)
"""
if not isinstance(box[0], float):
box = [float(box[i]) for i in range(len(box))]
result = []
left_anchor_num = int(
math.floor(
max(min(box[0], box[6]), 0) /
anchor_width)) # the left side anchor of the text box, downwards
right_anchor_num = int(
math.ceil(
min(max(box[2], box[4]), img.shape[1]) /
anchor_width)) # the right side anchor of the text box, upwards
# handle extreme case, the right side anchor may exceed the image width
if right_anchor_num * 16 + 15 > img.shape[1]:
right_anchor_num -= 1
# combine the left-side and the right-side x_coordinate of a text anchor into one pair
position_pair = [(i * anchor_width, (i + 1) * anchor_width - 1)
for i in range(left_anchor_num, right_anchor_num)]
y_top, y_bottom = cal_y_top_and_bottom(img, position_pair, box)
#print("image shape: %s, pair_num: %s, top_num:%s, bot_num:%s" % (img.shape, len(position_pair), len(y_top), len(y_bottom)))
for i in range(len(position_pair)):
position = int(position_pair[i][0] /
anchor_width) # the index of anchor box
h = y_bottom[i] - y_top[i] + 1 # the height of anchor box
cy = (float(y_bottom[i]) +
float(y_top[i])) / 2.0 # the center point of anchor box
result.append((position, cy, h))
draw_img_gt = draw_image.draw_box_h_and_c(draw_img_gt, position, cy, h)
draw_img_gt = draw_image.draw_box_4pt(draw_img_gt,
box,
color=(0, 0, 255),
thickness=1)
return result, draw_img_gt