ha4t.aircv.aircv 源代码

#!/usr/bin/env python
# -*- coding: utf-8 -*-
import os
import sys

import cv2
import numpy as np
from six import PY2, PY3

from ha4t.aircv.error import FileNotExistError
from ha4t.aircv.utils import cv2_2_pil, compress_image


[文档] def imread(filename, flatten=False): """根据图片路径,将图片读取为cv2的图片处理格式.""" if not os.path.isfile(filename): raise FileNotExistError("File not exist: %s" % filename) # choose image readin mode: cv2.IMREAD_UNCHANGED=-1, cv2.IMREAD_GRAYSCALE=0, cv2.IMREAD_COLOR=1, readin_mode = cv2.IMREAD_GRAYSCALE if flatten else cv2.IMREAD_COLOR if PY3: img = cv2.imdecode(np.fromfile(filename, dtype=np.uint8), readin_mode) else: filename = filename.encode(sys.getfilesystemencoding()) img = cv2.imread(filename, readin_mode) return img
[文档] def imwrite(filename, img, quality=10, max_size=None): """写出图片到本地路径,压缩""" if PY2: filename = filename.encode(sys.getfilesystemencoding()) pil_img = cv2_2_pil(img) compress_image(pil_img, filename, quality, max_size=max_size)
[文档] def show(img, title="show_img", test_flag=False): """在可缩放窗口里显示图片.""" cv2.namedWindow(title, cv2.WINDOW_NORMAL) cv2.imshow(title, img) if not test_flag: cv2.waitKey(0) cv2.destroyAllWindows()
[文档] def show_origin_size(img, title="image", test_flag=False): """原始尺寸窗口中显示图片.""" cv2.imshow(title, img) if not test_flag: cv2.waitKey(0) cv2.destroyAllWindows()
[文档] def rotate(img, angle=90, clockwise=True): """ 函数使图片可顺时针或逆时针旋转90、180、270度. 默认clockwise=True:顺时针旋转 """ def count_clock_rotate(img): # 逆时针旋转90° rows, cols = img.shape[:2] rotate_img = np.zeros((cols, rows)) rotate_img = cv2.transpose(img) rotate_img = cv2.flip(rotate_img, 0) return rotate_img # 将角度旋转转化为逆时针旋转90°的次数: counter_rotate_time = (4 - angle / 90) % 4 if clockwise else (angle / 90) % 4 for i in range(int(counter_rotate_time)): img = count_clock_rotate(img) return img
[文档] def crop_image(img, rect): """ 区域截图,同时返回截取结果 和 截取偏移; Crop image , rect = [x_min, y_min, x_max ,y_max]. (airtest中有用到) """ if isinstance(rect, (list, tuple)) and len(rect) == 4: height, width = img.shape[:2] # 获取在图像中的实际有效区域: x_min, y_min, x_max, y_max = [int(i) for i in rect] x_min, y_min = max(0, x_min), max(0, y_min) x_min, y_min = min(width - 1, x_min), min(height - 1, y_min) x_max, y_max = max(0, x_max), max(0, y_max) x_max, y_max = min(width - 1, x_max), min(height - 1, y_max) # 返回剪切的有效图像+左上角的偏移坐标: img_crop = img[y_min:y_max, x_min:x_max] return img_crop else: raise Exception("to crop a image, rect should be a list like: [x_min, y_min, x_max, y_max].")
[文档] def mark_point(img, point, circle=False, color=100, radius=20): """ 调试用的: 标记一个点 """ x, y = point # cv2.rectangle(img, (x, y), (x+10, y+10), 255, 1, lineType=cv2.CV_AA) if circle: cv2.circle(img, (x, y), radius, 255, thickness=2) cv2.line(img, (x - radius, y), (x + radius, y), color) # x line cv2.line(img, (x, y - radius), (x, y + radius), color) # y line return img
[文档] def mask_image(img, mask, color=(255, 255, 255), linewidth=-1): """ 将screen的mask矩形区域刷成白色gbr(255, 255, 255). 其中mask区域为: [x_min, y_min, x_max, y_max]. color: 顺序分别的blue-green-red通道. linewidth: 为-1时则完全填充填充,为正整数时为线框宽度. """ # 将划线边界外扩,保证线内区域不被线所遮挡: offset = int(linewidth / 2) return cv2.rectangle(img, (mask[0] - offset, mask[1] - offset), (mask[2] + linewidth, mask[3] + linewidth), color, linewidth)
[文档] def get_resolution(img): h, w = img.shape[:2] return w, h