我已经使用OCR API已有一段时间了,但是我想看看ognizeText API是否提高了性能。问题在于,调用ognizeText API只会返回一个空字符串。
如果我将API调用更改为/vision/v2.0/ocr,则一切正常。
import http.client, urllib.request, urllib.parse, urllib.error, base64
headers = {
# Request headers
'Content-Type': 'application/octet-stream',
'Ocp-Apim-Subscription-Key': 'MYAPIKEY',
}
params = urllib.parse.urlencode({
# Request parameters
'mode': 'Printed',
})
with open(filename, 'rb') as f:
data_image = f.read()
try:
conn = http.client.HTTPSConnection('canadacentral.api.cognitive.microsoft.com')
conn.request("POST", "/vision/v2.0/recognizeText?%s" % params, data_image, headers)
response = conn.getresponse()
data = response.read()
print(data)
conn.close()
except Exception as e:
print("[Errno {0}] {1}".format(e.errno, e.strerror))
答案 0 :(得分:0)
API仅返回202接受(如果您的请求有效)。这包括一个包含URL的标头 Operation-Location 。您随后对该URL发出了请求,以检索请求的状态/结果。