opencv-python学习(6)--六个图像处理小项目

发表于 更新于 分类于 ， 本文字数： 7.9k 阅读时长 ≈ 7 分钟

摘要

opencv的python图像处理学习笔记

转像素画、转字符画、图片九宫格、图片蒙太奇、幻影坦克、证件照裁剪。

部分代码参考互联网。

1.图像转像素画

效果如下

可莉像素

实现原理是通过将图片缩小10倍，然后用最邻近插值放大回原来的尺寸。

1
2
3
4
5
6
7
8
# 图片像素化
def pixel_image(self):
    img = self.image.copy()
    rows, cols = img.shape[:2]
    img = cv2.resize(img,(cols//10,rows//10))
    img = cv2.resize(img,(cols,rows),interpolation=cv2.INTER_NEAREST)
    win_name = "pixel"
    UIFunctions.wait_key(self,img,win_name)

其中wait_key函数内容是用opencv打开图片，然后接受一些保存等指令的函数。

2.图片转字符画

效果图

字符画

原理是把图片转为灰度，然后根据灰度值将像素替换为对应字重的字符最后保存为txt

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
# 转为字符画
def character_image(self):
    img = UIFunctions.trun_gray(self,self.image)
    rows, cols = img.shape[:2]
    rows = rows//2
    img = cv2.resize(img,(cols,rows))
    lstChars = self.config.get('General','file_character_list')
    output = self.config.get('General','file_save_path') + "\\character.txt"
    g2c = 256/len(lstChars)
    
    txt = ""
    for i in range(rows):
        for j in range(cols):
            txt += lstChars[int(img[i,j]/g2c)]
        txt += '\n'

    with open(output, 'w') as f:
        f.write(txt)
    QMessageBox.information(None, '提示', 'character.txt输出成功',QMessageBox.Ok)

为了表现效果，这里将高缩了一半。output是文件路径，lstChars是自定义的字符。上图使用的字符串是

1
$B8&WM*oahkbdpqwmZO0QLCJUYXzcvunxrjft/\|1?-_+~<>i!lI,^`.

注意把字重大的放前面。

3.图片九宫格

图片九宫格是发朋友圈常用的一种格式，该项目可以辅助将一张图片切割成9张。

效果

图片九宫格

这个没啥说的就是切一下图片,因为懒得pip没用PIL库就直接对数组操作了。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
# 九宫格
def ninegrid_image(self):
    # 切片函数
    def cut_image(img,width):
        box_list = []
        if len(img.shape) == 2:
            for i in range(0, 3):
                for j in range(0, 3):
                    box = img[i * width:(i+1) * width, j * width : (j + 1) * width]
                    box_list.append(box)
        elif len(img.shape) == 3:
            for i in range(0, 3):
                for j in range(0, 3):
                    box = img[i * width:(i+1) * width, j * width : (j + 1) * width,:]
                    box_list.append(box)
        return box_list

    img = self.image.copy()
    value, ok = QInputDialog.getInt(self, "输入窗口", "请输入裁剪后的边值像素(正方形):", 300, 0, 2000, 100)
    if ok:
        value = int(value)*3
        img = cv2.resize(img,(value,value))
        box_list = cut_image(img,value//3)
        for i,box in enumerate(box_list):
            save_name = self.config.get('General','file_save_path') + '\\ninegrid_image\\' + str(i+1)+'.png'
            cv2.imwrite(save_name,box)
        QMessageBox.information(None, '提示', 'ninegrid_image输出成功',QMessageBox.Ok)

4.图片蒙太奇

图片蒙太奇是指用一组图片去组合成一张大图的艺术效果。

图片有点大，放本地打开后再截图的图。

蒙太奇

代码主要参考CSDN蒙太奇照片制作(Opencv+Python)

原理是缩放待处理图片，然后匹配区块和图片的直方图找到最合适的，最后做一个融合让效果变好些。

理论上图库内容丰富越好。

5.幻影坦克

幻影坦克是现代互联网图片表示的艺术，缩略图上看是一张图，但实际点开就变成了另一张，常见于贴吧。

效果图

幻影坦克

原理参考b站小迷糊老师的视频《幻影坦克基础版》

简单讲就是利用背景白色和黑色来叠加出不一样的图片。

灰度处理方面借用了《基于OpenCV实现的灰度图幻影坦克》一文的思路

代码 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
# 幻影坦克
def mirage_tank_image(self):
    inside_pic = np.uint8(self.image.copy() * 0.35)
    h,w = inside_pic.shape[:2]
    get_filename_path, ok = QFileDialog.getOpenFileName(self,"选取表图",self.config.get('General','file_open_path'))
    if ok:
        out_pic = cv2.imread(get_filename_path)
        # 调整表图大小 与里图一致
        out_pic = cv2.resize(out_pic,(w,h))
        # 如果强度小于100 拉到100
        out_pic[out_pic<100] = 100
        # 转变为灰度图
        inside_pic = UIFunctions.trun_gray(self,inside_pic)
        out_pic = UIFunctions.trun_gray(self,out_pic)
        # 建立新的画布 
        new_pic = np.zeros((h,w,4), np.uint8)
        # print(new_pic.shape)
        # 遍历图片 套用公式
        for i in range(h):
            for j in range(w):
                alpha = 255 - (out_pic[i,j]-inside_pic[i,j])
                if alpha == 0:
                    alpha = 1
                p_new = np.uint8(255*inside_pic[i,j]/alpha)
                new_pic[i,j,0] = p_new
                new_pic[i,j,1] = p_new
                new_pic[i,j,2] = p_new
                new_pic[i,j,3] = alpha
        win_name = "tank"
        UIFunctions.wait_key(self,new_pic,win_name)

6.证件照处理

如果你对证件照态度比较随便的话，可以直接用手机背靠墙拍一张，然后用这个方法处理，直接用打印机打印出来。

效果图

证件照

原理是建立对应大小的画布，然后把图片处理成1或2寸的，贴到对应位置。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
# 证件照处理
def certificate_image(self):
    # 1寸大小
    WIDTH_1IN = 295
    HEIGHT_1IN = 413
    # 2寸大小
    WIDTH_2IN = 413
    HEIGHT_2IN = 626
    # 5寸大小
    WIDTH_5IN = 1500
    HEIGHT_5IN = 1050
    # 非全景6寸照片
    WIDTH_6IN = 1950
    HEIGHT_6IN = 1300
    def cut_photo(photo,choice):
        """
        将照片按照比例进行裁剪并缩放成1寸、2寸
        :param photo: 待处理的照片
        :param choice: <int> 1代表1寸，2代表2寸
        :return: 处理后的照片
        """
        h,w = photo.shape[:2]
        rate = h / w
        if choice == 1:
            if rate < (HEIGHT_1IN/WIDTH_1IN):
                cut = int((w - int(h / HEIGHT_1IN * WIDTH_1IN)) / 2)
                cut_p = photo[:,cut:w-cut,:]
            else:
                cut = int((h - int(w / WIDTH_1IN * HEIGHT_1IN)) / 2)
                cut_p = photo[cut:w-cut,:,:]
            return cv2.resize(cut_p, (WIDTH_1IN, HEIGHT_1IN))
        if choice == 2:
            if rate < (HEIGHT_2IN/WIDTH_2IN):
                cut = int((w - int(h / HEIGHT_2IN * WIDTH_2IN)) / 2)
                cut_p = photo[:,cut:w-cut,:]
            else:
                cut = int((h - int(w / WIDTH_2IN * HEIGHT_2IN)) / 2)
                cut_p = photo[cut:w-cut,:,:]
            return cv2.resize(cut_p, (WIDTH_2IN, HEIGHT_2IN))

    # 5寸照排版1
    def layout_photo_5_1(photo):
        photo = cut_photo(photo,1)
        img = np.ones((HEIGHT_5IN,WIDTH_5IN, 3), np.uint8)*255
        img[HEIGHT_5IN//2,:,:] = 128            # 横线
        img[:,int(WIDTH_5IN*0.25),:] = 128      # 竖线1
        img[:,int(WIDTH_5IN*0.5),:] = 128      # 竖线2
        img[:,int(WIDTH_5IN*0.75),:] = 128      # 竖线3
        # 确定位置放上图片
        focus_point = [0.125 * WIDTH_5IN,0.25 * HEIGHT_5IN]
        start_point = [focus_point[0] - 0.5 * WIDTH_1IN, focus_point[1] - 0.5 * HEIGHT_1IN]
        end_point = [focus_point[0] + 0.5 * WIDTH_1IN, focus_point[1] + 0.5 * HEIGHT_1IN]
        for i in range(0,2):
            for j in range(0,4):
                img[int(start_point[1]+0.5*i*HEIGHT_5IN):int(end_point[1]+0.5*i*HEIGHT_5IN),int(start_point[0]+j*WIDTH_5IN/4):int(end_point[0]+j*WIDTH_5IN/4),:] = photo
        return img
    # 5寸混合排版2
    def layout_photo_5_mix(photo):
        photo1 = cut_photo(photo,1)
        photo2 = cut_photo(photo,2)
        photo2 = np.rot90(photo2)
        img = np.ones((HEIGHT_5IN,WIDTH_5IN, 3), np.uint8)*255
        img[HEIGHT_5IN//2,:,:] = 128            # 横线
        img[:,int(WIDTH_5IN*0.25),:] = 128      # 竖线1
        img[:,int(WIDTH_5IN*0.5),:] = 128      # 竖线2
        # 确定位置放上1寸图片
        focus_point = [0.125 * WIDTH_5IN,0.25 * HEIGHT_5IN]
        start_point = [focus_point[0] - 0.5 * WIDTH_1IN, focus_point[1] - 0.5 * HEIGHT_1IN]
        end_point = [focus_point[0] + 0.5 * WIDTH_1IN, focus_point[1] + 0.5 * HEIGHT_1IN]
        for i in range(0,2):
            for j in range(0,2):
                img[int(start_point[1]+0.5*i*HEIGHT_5IN):int(end_point[1]+0.5*i*HEIGHT_5IN),int(start_point[0]+j*WIDTH_5IN/4):int(end_point[0]+j*WIDTH_5IN/4),:] = photo1
        # 确定位置放上2寸图片
        focus_point2 = [0.75 * WIDTH_5IN, 0.25 * HEIGHT_5IN]
        start_point2 = [focus_point2[0] - 0.5 * HEIGHT_2IN, focus_point2[1] - 0.5 * WIDTH_2IN]
        end_point2 = [focus_point2[0] + 0.5 * HEIGHT_2IN, focus_point2[1] + 0.5 * WIDTH_2IN]
        img[int(start_point2[1]):int(end_point2[1]),int(start_point2[0]):int(end_point2[0]),:] = photo2
        img[int(start_point2[1]+ 0.5 * HEIGHT_5IN):int(end_point2[1]+ 0.5 * HEIGHT_5IN),int(start_point2[0]):int(end_point2[0]),:] = photo2
        return img
    # 6寸混合排版1
    def layout_photo_6_mix(photo):
        photo1 = cut_photo(photo,1)
        photo2 = cut_photo(photo,2)
        photo1 = np.rot90(photo1)
        img = np.ones((HEIGHT_6IN,WIDTH_6IN, 3), np.uint8)*255
        img[HEIGHT_6IN//2,:,:] = 128            # 横线1
        img[HEIGHT_6IN//4,:WIDTH_6IN//2,:] = 128 # 横线2
        img[int(HEIGHT_6IN*0.75),:WIDTH_6IN//2,:] = 128 # 横线2
        img[:,int(WIDTH_6IN*0.25),:] = 128      # 竖线1
        img[:,int(WIDTH_6IN*0.5),:] = 128      # 竖线2
        img[:,int(WIDTH_6IN*0.75),:] = 128      # 竖线3
        # 确定位置放上1寸图片
        focus_point = [0.125 * WIDTH_6IN, 0.125 * HEIGHT_6IN]
        start_point = [focus_point[0] - 0.5 * HEIGHT_1IN, focus_point[1] - 0.5 * WIDTH_1IN]
        end_point = [focus_point[0] + 0.5 * HEIGHT_1IN, focus_point[1] + 0.5 * WIDTH_1IN]
        for i in range(0,4):
            for j in range(0,2):
                img[int(start_point[1]+0.25*i*HEIGHT_6IN):int(end_point[1]+0.25*i*HEIGHT_6IN),int(start_point[0]+j*WIDTH_6IN/4):int(end_point[0]+j*WIDTH_6IN/4),:] = photo1
        # 确定位置放上2寸图片
        focus_point2 = [0.625 * WIDTH_6IN, 0.25 * HEIGHT_6IN]
        start_point2 = [focus_point2[0] - 0.5 * WIDTH_2IN, focus_point2[1] - 0.5 * HEIGHT_2IN]
        end_point2 = [focus_point2[0] + 0.5 * WIDTH_2IN, focus_point2[1] + 0.5 * HEIGHT_2IN]
        for i in range(0,2):
            for j in range(0,2):
                img[int(start_point2[1]+0.5*i*HEIGHT_6IN):int(end_point2[1]+0.5*i*HEIGHT_6IN),int(start_point2[0]+j*WIDTH_6IN/4):int(end_point2[0]+j*WIDTH_6IN/4),:] = photo2

        return img

    img = self.image.copy()
    items = ('1寸', '2寸', '5寸(8*1寸)', '5寸(4*1寸+2*2寸)', '6寸(8*1寸+4*2寸)')
    item,ok = QInputDialog.getItem(self,"区域分割","选择需要分成几块",items,3,False)
    if ok:
        if item == '1寸':
            img = cut_photo(img,1)
        elif item == '2寸':
            img = cut_photo(img,2)
        elif item == '5寸(8*1寸)':
            img = layout_photo_5_1(img)
        elif item == '5寸(4*1寸+2*2寸)':
            img = layout_photo_5_mix(img)
        elif item == '6寸(8*1寸+4*2寸)':
            img = layout_photo_6_mix(img)

        win_name = "ID photo"
        UIFunctions.wait_key(self,img,win_name)