How to use _cal_confidence method in Airtest

...44        rect, matches, good = self.get_rect_from_good_matches(im_source, im_search, kp_sch, des_sch, kp_src, des_src)45        if not rect:46            return None47        # 第三步, 从匹配图片截取矩阵范围,并缩放到模板大小,进行模板匹配求出相似度48        confidence = self._cal_confidence(im_source=im_source, im_search=im_search, rect=rect, rgb=rgb)49        best_match = generate_result(rect=rect, confi=confidence)50        return best_match if confidence > threshold else None51    @print_all_result52    def find_all(self, im_source, im_search, threshold: Union[int, float] = None,53                 max_count: int = 10,54                 rgb: bool = None):55        threshold = threshold or self.threshold56        rgb = rgb is None and self.rgb or rgb57        im_source, im_search = self.check_detection_input(im_source, im_search)58        if not im_source or not im_search:59            return None60        result = []61        # 第一步: 获取特征点集62        kp_sch, des_sch = self.get_keypoints_and_descriptors(image=im_search.rgb_2_gray())63        kp_src, des_src = self.get_keypoints_and_descriptors(image=im_source.rgb_2_gray())64        while len(kp_src) > 2 or len(kp_sch) > 2:65            rect, matches, good = self.get_rect_from_good_matches(im_source, im_search,66                                                                  kp_sch, des_sch,67                                                                  kp_src, des_src)68            if not rect:69                break70            confidence = self._cal_confidence(im_source=im_source, im_search=im_search, rect=rect, rgb=rgb)71            if confidence > threshold and len(result) < max_count:72                result.append(generate_result(rect, confidence))73            kp_src, des_src = self.delect_good_descriptors(good, kp_src, des_src)74            kp_src, des_src = self.delect_rect_descriptors(rect, kp_src, des_src)75        return result76    def _cal_confidence(self, im_source, im_search, rect: Rect, rgb: bool):77        """ 将截图和识别结果缩放到大小一致,并计算可信度 """78        try:79            target_img = im_source.crop_image(rect)80        except OverflowError:81            raise MatchResultError("Target area({}) out of screen{}".format(rect, im_source.size))82        h, w = im_search.size83        target_img.resize(w, h)84        if rgb:85            confidence = self.template.cal_rgb_confidence(im_source=im_search, im_search=target_img)86        else:87            confidence = self.template.cal_ccoeff_confidence(im_source=im_search, im_search=target_img)88        confidence = (1 + confidence) / 289        return confidence90    @staticmethod...

...59def _dist(point, target):60    x, y = point61    m, n = target62    return np.sqrt((x - m) ** 2 + (y - n) ** 2)63def _cal_confidence(dist):64    # linear function. score decreases as the distance increases65    return -CONFIDENCE_RATE * dist + 166def is_id_card(crop_box, face_box, debug=False):67    xleft, xright, xtop, xbottom = crop_box68    yleft, yright, ytop, ybottom = face_box69    yleft += xleft70    yright += xleft71    ytop += xtop72    ybottom += xtop73    if debug:74        print('crop box: ', xleft, xright, xtop, xbottom)75        print('face box: ', yleft, yright, ytop, ybottom)76    crop_mid = _calc_mid(xleft, xright, xtop, xbottom)77    face_mid = _calc_mid(yleft, yright, ytop, ybottom)78    target = (xright + crop_mid[0]) // 2, crop_mid[1]79    return _cal_confidence(_dist(face_mid, target))80def id_card_detect(img, args):81    sess, detection_graph, category_index = args82    tensors = get_tensors(detection_graph)83    try:84        # crop85        cropped_img, crop_box, score = crop_id(img, sess, tensors, category_index)86    except:87        import traceback88        traceback.print_exc()89        return None, 090    return cropped_img if cropped_img is not None else None, score91if __name__ == "__main__":92    img = cv.imread('test_samples/1121.png')93    args = load_model()...