45 lines
1.2 KiB
Python
45 lines
1.2 KiB
Python
|
import cv2
|
||
|
|
import time
|
||
|
|
from pp_onnx.onnx_paddleocr import ONNXPaddleOcr, draw_ocr
|
||
|
|
|
||
|
|
# 优化参数配置，提高速度
|
||
|
|
model = ONNXPaddleOcr(
|
||
|
|
use_angle_cls=True,
|
||
|
|
use_gpu=False,
|
||
|
|
providers=['AzureExecutionProvider'],
|
||
|
|
provider_options=[{'device_id': 0}],
|
||
|
|
det_limit_type='max',
|
||
|
|
)
|
||
|
|
|
||
|
|
def sav2Img(org_img, result, name="./result_img/3.jpg"):
|
||
|
|
from PIL import Image
|
||
|
|
result = result[0]
|
||
|
|
image = org_img[:, :, ::-1]
|
||
|
|
boxes = [line[0] for line in result]
|
||
|
|
txts = [line[1][0] for line in result]
|
||
|
|
scores = [line[1][1] for line in result]
|
||
|
|
im_show = draw_ocr(image, boxes, txts, scores)
|
||
|
|
im_show = Image.fromarray(im_show)
|
||
|
|
im_show.save(name)
|
||
|
|
|
||
|
|
|
||
|
|
# 执行OCR推理
|
||
|
|
img = cv2.imread('./test_img/3.jpg')
|
||
|
|
if img is None:
|
||
|
|
print(f"❌ 未找到图像文件")
|
||
|
|
else:
|
||
|
|
# 图像预处理 - 缩小图像尺寸加速处理
|
||
|
|
h, w = img.shape[:2]
|
||
|
|
max_size = 1080
|
||
|
|
if max(h, w) > max_size:
|
||
|
|
scale = max_size / max(h, w)
|
||
|
|
img = cv2.resize(img, (int(w * scale), int(h * scale)))
|
||
|
|
|
||
|
|
s = time.time()
|
||
|
|
result = model.ocr(img)
|
||
|
|
e = time.time()
|
||
|
|
print(f"total time: {e - s:.3f} 秒")
|
||
|
|
print("result:", result)
|
||
|
|
for box in result[0]:
|
||
|
|
print(box)
|
||
|
|
sav2Img(img, result)
|