mirror of
https://github.com/nizhangdaye/TLD_Code.git
synced 2024-11-23 20:39:06 +08:00
90 lines
3.1 KiB
Python
90 lines
3.1 KiB
Python
import cv2 as cv
|
||
import numpy as np
|
||
from matplotlib import pyplot as plt
|
||
import number_processing
|
||
|
||
|
||
def cv_show(name, img):
|
||
"""定义cv_show函数，用于显示图像"""
|
||
cv.namedWindow(name, cv.WINDOW_NORMAL) # 窗口大小可调
|
||
cv.resizeWindow(name, img.shape[1], img.shape[0]) # 窗口大小可调
|
||
cv.imshow(name, img)
|
||
cv.waitKey(0)
|
||
cv.destroyAllWindows()
|
||
|
||
|
||
def cv_info(name, img):
|
||
"""
|
||
定义cv_info函数，确认图像是否成功读取，并且打印图像的基本信息
|
||
"""
|
||
print(f"图像 {name} 信息如下：")
|
||
if img is None:
|
||
print("无法读取图像")
|
||
else:
|
||
print("图像数据类型：", img.dtype)
|
||
print("图像类型：", type(img))
|
||
print("图像尺寸：", img.shape)
|
||
print("图像行数：", img.shape[0])
|
||
print("图像列数：", img.shape[1])
|
||
print()
|
||
|
||
|
||
def find_line(img):
|
||
"""
|
||
定义find_line函数，用于寻找分界线
|
||
返回一段图像的宽度
|
||
"""
|
||
for i in range(img.shape[1]): # 遍历 a 列
|
||
if np.all(img[0, i] >= [230, 230, 230]):
|
||
# 求该列像素均值，判断是否为白色
|
||
mean = np.mean(img[:, i])
|
||
if mean >= 240: # 像素均值大于 230 判断为白色
|
||
print(f"分界线位置：{i}, 像素均值：{mean}", end="\n\n")
|
||
return i
|
||
return -1 # 如果没有找到符合条件的分界线，则返回图像的宽度
|
||
|
||
|
||
def split_img(img):
|
||
"""
|
||
定义split_img函数，用于将图像分割并保存
|
||
无返回值
|
||
"""
|
||
img_width = find_line(img) # 返回图像的宽度
|
||
n = 0 # 用于计数分割得到的图像数
|
||
for i in range(0, img.shape[1], img_width + 1): # 遍历 img 的每一列
|
||
image = img[:, i: i + img_width]
|
||
n += 1
|
||
cv.imwrite(f"data/saved_img/image_{n}.PNG", image) # 保存图片 无损压缩格式 PNG
|
||
|
||
|
||
def split_num(img):
|
||
"""
|
||
定义split_num函数，用于分割数字并保存
|
||
返回数字
|
||
"""
|
||
# 将彩色图像转化为灰度图像
|
||
img_gray = cv.cvtColor(img, cv.COLOR_BGR2GRAY)
|
||
# 二值化处理，你可以根据需要调整阈值
|
||
_, thresh = cv.threshold(img_gray, 128, 255, cv.THRESH_BINARY_INV)
|
||
# 查找轮廓
|
||
contours, _ = cv.findContours(thresh, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)
|
||
# 对轮廓进行排序，以便从左到右处理数字（可选）
|
||
contours = sorted(contours, key=lambda x: cv.boundingRect(x)[0])
|
||
|
||
# 用于存储识别出的数字字符串
|
||
recognized_digits = []
|
||
|
||
# 遍历每个轮廓，并保存每个数字的图片
|
||
for i, contour in enumerate(contours):
|
||
x, y, w, h = cv.boundingRect(contour)
|
||
digit = thresh[y:y + h, x:x + w]
|
||
# print(f"第{i+1}个数字的尺寸：{digit.shape[0]}")
|
||
if digit.shape[0] < 5: # 小数点的轮廓太小，跳过
|
||
break
|
||
|
||
recognized_digits.append(str(number_processing.num_list.index(digit.tolist())))
|
||
|
||
# # 保存数字图片
|
||
# cv.imwrite(f'data/saved_img/digit_{i}.png', digit)
|
||
return int("".join(recognized_digits))
|