最近终于找到一个好的方法，使用Python的OpenCV模块识别滑动验证码的缺口，可以将滑动验证码中的缺口识别出来了。
测试使用如下两张图片：
target.jpg
template.png
现在想要通过“template.png”在“target.jpg”中找到对应的缺口，代码实现如下：
# encoding=utf8 import cv2 import numpy as np def show(name): cv2.imshow('Show', name) cv2.waitKey(0) cv2.destroyAllWindows() def main(): otemp = 'template.png' oblk = 'target.jpg' target = cv2.imread(otemp, 0) template = cv2.imread(oblk, 0) w, h = target.shape[::-1] temp = 'temp.jpg' targ = 'targ.jpg' cv2.imwrite(temp, template) cv2.imwrite(targ, target) target = cv2.imread(targ) target = cv2.cvtColor(target, cv2.COLOR_BGR2GRAY) target = abs(255 - target) cv2.imwrite(targ, target) target = cv2.imread(targ) template = cv2.imread(temp) result = cv2.matchTemplate(target, template, cv2.TM_CCOEFF_NORMED) x, y = np.unravel_index(result.argmax(), result.shape) # 展示圈出来的区域 cv2.rectangle(template, (y, x), (y + w, x + h), (7, 249, 151), 2) show(template) if __name__ == '__main__': main()