Skip to main content
 首页 » 编程设计

python之Pythons PIL (v2.6) 中的 TypeError 消息 : integer expected, float

2025年05月04日8jirigala

第一次来论坛。希望我足够具体。 在 PIL 中使用 ImageChops,我试图将两个图像相乘(都是 mode="L"),但我总是收到相同的错误消息。我到处都看过,但找不到任何有用的东西。我将不胜感激任何有用的想法! 附上代码的相关部分。

    def point(self, f, searchImage, technique): # technique - inpaint or bicubic 
 
    dimx, dimy = searchImage.size 
 
    reader = csv.reader(f) 
    for line in reader: #f.readlines(): 
        coord = line 
        print coord 
        if searchImage.size[0] > float(coord[0])+95.5 and searchImage.size[1]\ 
           > float(coord[1])+95.5:  
            box = (float(coord[0])-93.5,float(coord[1])-93.5,\ 
                   float(coord[0])+95.5,float(coord[1])+95.5)     # left upper right 
        elif searchImage.size[0] < float(coord[0])+95.5 and searchImage.size[1]\ 
             > float(coord[1])+95.5: 
            box = (float(coord[0])-93.5,float(coord[1])-93.5,\ 
                   searchImage.size[0]-0.5,float(coord[1])+95.5)  # size of box 
            # depends on pixel size. A pixel size of 14 micrometer results in a 
            # cross size of 189 pixels 
        else: 
            box = (float(coord[0])-93.5,float(coord[1])-93.5,\ 
                   float(coord[0])+95.5,searchImage.size[1]-0.5) 
 
        box = (math.floor(box[0]), math.floor(box[1]), math.floor(box[2]),\ 
               math.floor(box[3]))  
 
        searchCrop = searchImage.crop(box) 
 
        c_x = int(float(coord[1]))  
        c_y = int(float(coord[0]))  
        abst_y = c_x - int(math.floor(box[1])) - 1 # x shift 
 
        center = num.asarray(searchImage)[c_x,c_y] 
        if center == 0: 
            center = center + 0.00001 # to avoid division by zero 
        val = [num.asarray(searchImage)[c_x-1,c_y+1], num.asarray(searchImage)\ 
               [c_x-1,c_y-1], num.asarray(searchImage)[c_x+1,c_y-1], \ 
               num.asarray(searchImage)[c_x+1,c_y+1]] # ERDAS upper right, 
        # upper left, lower left, lower right 
 
        val_dict = {0:1,1:-1,2:-1,3:1} 
        flag = val_dict[val.index(min(val))] 
        if float(min(val))/center > 2. or min(val) > 100: 
            flag = 0 
 
        newima = num.zeros( (searchCrop.size[1], searchCrop.size[0]),\ 
                            dtype = "float") 
 
        Ayo = num.array(int(searchCrop.size[0])*[255])  
        Ay = num.array((abst_y + flag)*[255] + 3*[0] + ((int(searchCrop.size[0]\ 
                                                             )-3-abst_y)-flag)*[255])  
        Ax = num.array(int(searchCrop.size[0])*[0])   
        Kx = num.array(3*[Ayo] + ((int(searchCrop.size[1])-9)/2+flag)*[Ay] + 3*[Ax] \ 
                       + ((int(searchCrop.size[1])-9)/2-flag)*[Ay] + 3*[Ayo])    
 
        Kxlist = list(itertools.chain(*Kx)) 
 
        i=0 
        for y in range(int(searchCrop.size[1])):  
            for x in range(int(searchCrop.size[0])):         
                newima[y,x] = Kxlist[i+y+x] 
            i=i+x 
 
        kernel = Image.fromarray(newima) 
        kernel = kernel.convert(mode="L") 
 
        # ----- 
        modified = ImageChops.multiply(searchCrop,kernel)   # Results in an image  
        # where the pixels along the cross axes will get a value of 0 
        # --- 

错误信息如下:

File "D:\GIS_dbase\Data\hma_cci\hexagon\KH9_Python\interpolate_cross.py", line 58, in  
 crossInterpolation filledImage = self.p_model.point(f, searchImage, method)  
File "D:\GIS_dbase\Data\hma_cci\hexagon\KH9_Python\interpolate_cross.py", line 207, in  
 point modified = ImageChops.multiply(searchCrop,kernel)   # Results in an image where  
 the pixels along the cross axes will get a value of 0 
File "C:\Python27\lib\site-packages\PIL\ImageChops.py", line 119, in multiply 
 image1.load() 
File "C:\Python27\lib\site-packages\PIL\Image.py", line 1730, in load 
 self.im = self.im.crop(self.__crop) 
TypeError: integer argument expected, got float 

请您参考如下方法:

问题是 PIL 的 crop 方法采用一个包含 4 个整数值的元组,但您传递的是 float 。这应该有效:

box = tuple([int(math.floor(x)) for x in box])