如何在屏幕上找到具有透明背景的图像？

Question

所以，我有一张具有透明背景和尺寸（Width_1、High_1）的图像，而我要做的是在屏幕上找到这张图像（但不必具有相同的尺寸）比方说尺寸（Width_2，High_2），所以考虑到这一点，我编写了以下代码：

def resize_untill_find(image,bigger_or_smaller, min_size = 50 , max_size = 150 , one_match_per_point = 'on'): #(imagem, diminuir ou aumentar, min_size = reduce the image until min_size%, max_size = expand the image until max_size%)
    print('Press CTRL+ALT to start:')
    while True:
        if keyboard.is_pressed('control+alt'):
            break
    image_info = Ferramentas.get_image_first_info(image)# get width, high, and pixels number off the image
    image = cv2.imread(image)
    image_rate = image_info['whidth'] / image_info['high'] #proportion Width/High
    ocurrencies = {} # [x,y,w,h]
    match_number = 1 # number of the ocurrency
    high = image_info['high']
    if bigger_or_smaller == 'smaller':
        while int((high/image_info['high'])*100) >= int(min_size): # while the resize are not under min_size
            print('{}%'.format(round((high/image_info['high'])*100)))
            width = round(image_info['whidth']-((image_info['high'] - high) * image_rate))
            dimension = (width,high) # new dimension of the image
            print('Dimension:', dimension)
            image_resized = cv2.resize(image,dimension) #resize the image
                matches_position = pyautogui.locateAllOnScreen(image_resized, confidence = 0.9)
                for region in matches_position:
                    region = list(region)
                    ocurrencies['Match_{}'.format(match_number)] = region # region = [X,Y, new_Width , new_High]
                    match_number += 1 
            if one_match_per_point == 'on':
                match_position = pyautogui.locateOnScreen(image_resized, confidence = 0.9)
                if match_position != None:
                    match_position = list(match_position)
                    ocurrencies['Match_{}'.format(match_number)] = match_position # region = [X,Y, new_Width , new_High]
                    match_number += 1
            high -= 1
    if bigger_or_smaller == 'bigger':
        while int((high/image_info['high'])*100) <= int(max_size):
            print('{}%'.format(round((high/image_info['high'])*100)))
            width = round(image_info['whidth']+((high - image_info['high']) * image_rate))
            dimension = (width,high) # image new dimension
            print('Dimension:', dimension)
            image_resized = cv2.resize(image,dimension) #resize the imge with new dimensions
            if one_match_per_point == 'off':
                matches_position = pyautogui.locateAllOnScreen(image_resized, confidence = 0.9) # lista todos os matches 90% iguais ao da imagem redimensionada
                for region in matches_position:
                    region = list(region)
                    ocurrencies['Match_{}'.format(match_number)] = region # region = [X,Y, new_Width , new_High]
                    match_number += 1   
            if one_match_per_point == 'on':
                match_position = pyautogui.locateOnScreen(image_resized, confidence = 0.9)
                if match_position != None:
                    match_position = list(match_position)
                    ocurrencies['Match_{}'.format(match_number)] = match_position # region = [X,Y, new_Width , new_High]
                    match_number += 1 
            high += 1
    print(ocurrencies)
    cv2.waitKey(0)
    return ocurrencies

我不知道是否有更好的方法来做到这一点，如果有，请告诉我，我很乐意改进代码（必须尽可能快），但主要问题是：它无法匹配具有透明度的模板，以及透明度渐变背景。

所以我有这张图片：

并且程序无法找到图像并获取它的坐标。

*编辑：

所以我可以拍下这张照片：

并调整大小，直到在此图像上找到匹配项：

但是当我尝试拍摄具有透明背景的 .png 图片时，它不起作用：

（模板）并尝试在这个上找到匹配项 (screen1) 或这个(screen2)。

您可以注意到第一个模板的尺寸与 screen1 相同，但背景为深灰色，而 screen2 的背景也不同且较小（但保持原来的Width/High比例）比原来的。

所以我想要的是能够使用 template 在屏幕的某个位置找到 screen1 或 screen2。

Answer 1

def search(screen, img):
sx, sy = screen.size
ix, iy = img.size
for xstart in range(sx - ix): 
    for ystart in range(sy - iy):
        #search for the pixel on the screen that equals the pixel at img[0:0]
        if img.getpixel((0,0)) == screen.getpixel((xstart, ystart)):
            match = 1 #temporary
            for x in range(ix): #check if first row of img is on this coords
                if img.getpixel((x,0)) <> screen.getpixel((xstart+x, ystart)):
                    match = 0 #if there's any difference, exit the loop
                    break 
            if match == 1: #otherwise, if this coords matches the first row of img
                for x in range(ix): 
                    for y in range(iy):
                        #check every pixel of the img
                        if img.getpixel((x,y)) <> screen.getpixel((xstart+x, ystart+y)):
                            match = 0 #any difference - break
                            break
                if match == 1: return (xstart, ystart) #return top-left corner coordinates
return (-1,-1) #or this, if not found
#i found this somewhere else