使用pytesseract在一串单词周围画一个矩形
使用pytesseract在一串单词周围画一个矩形
提问于 2022-10-27 11:03:42
这是我的形象：
我能认出以下几个字：
我需要检查图像中是否有带有特定文本的一行，并用矩形突出显示这一行。例如。我检查是否有“时间，这是最坏的”。然后我希望看到：
我怎样才能做到这一点？
我的代码：
import cv2 as cv
from pytesseract import pytesseract, Output
unchanged_image = cv.imread("i1Abv.png", cv.IMREAD_UNCHANGED)
initial_image = cv.imread("i1Abv.png", 0)
cv.imshow('', initial_image)
cv.waitKey(0)
ret, image = cv.threshold(initial_image, 100, 255, cv.THRESH_BINARY)
cv.imshow('', image)
cv.waitKey(0)
results = pytesseract.image_to_data(image, output_type=Output.DICT, config="--psm 6")
for i in range(0, len(results["text"])):
# extract the bounding box coordinates of the text region from
# the current result
x = results["left"][i]
y = results["top"][i]
w = results["width"][i]
h = results["height"][i]
# extract the OCR text itself along with the confidence of the
# text localization
text = results["text"][i]
conf = float(results["conf"][i])
# filter out weak confidence text localizations
if conf > 10:
# strip out non-ASCII text, so we can draw the text on the image
# using OpenCV, then draw a bounding box around the text along
# with the text itself
cv.rectangle(unchanged_image, (x, y), (x + w, y + h), (0, 255, 0), 2)
cv.putText(img=unchanged_image, text=text, org=(x, y), fontFace=cv.FONT_HERSHEY_COMPLEX,
fontScale=0.3, color=(36, 0, 255), thickness=1)
cv.imshow('', unchanged_image)
cv.waitKey(0)回答 1
Stack Overflow用户
回答已采纳
发布于 2022-10-27 19:00:02
pytesseract.image_to_data为每个可识别的文本块提供line_num。您可以将所有识别信息按line_num分组，然后将单词连接到一个文本行中。此外，您还需要找到一个边框，其中包括每个单词的所有文本框(您可以使用left、top、width、height获得它们)。你可以通过cv2.boundingRect找到它
结果：
代码：
import cv2
import numpy as np
import pytesseract
original_image = cv2.imread("text_line.png")
gray = cv2.cvtColor(original_image, cv2.COLOR_BGR2GRAY)
target_text = "times, it was the worst"
df = pytesseract.image_to_data(gray, lang="eng", config="--psm 6", output_type=pytesseract.Output.DATAFRAME)
# group recognized words by lines
for line_num, words_per_line in df.groupby("line_num"):
# filter out words with a low confidence
words_per_line = words_per_line[words_per_line["conf"] >= 5]
if not len(words_per_line):
continue
words = words_per_line["text"].values
line = " ".join(words)
print(f"{line_num} '{line}'")
if target_text in line:
print("Found a line with specified text")
word_boxes = []
for left, top, width, height in words_per_line[["left", "top", "width", "height"]].values:
word_boxes.append((left, top))
word_boxes.append((left + width, top + height))
x, y, w, h = cv2.boundingRect(np.array(word_boxes))
cv2.rectangle(original_image, (x, y), (x + w, y + h), color=(255, 0, 255), thickness=3)
cv2.imwrite("out.jpg", original_image)页面原文内容由Stack Overflow提供。腾讯云小微IT领域专用引擎提供翻译支持
原文链接：
https://stackoverflow.com/questions/74221064
复制相关文章
相似问题
腾讯云开发者
