同一张图片,不同拍照。清晰度差不多,为什么有时能识别,有时乱码
这是调用API接口的问题?
您好,该问题已经解决,预计5个工作日内后端升级解决,前段无感知升级,您可以在10.26日重新测试线上接口。
该问题目前是badcase,已经优化,估计马上上线。
你可以识别同一张图片的同一参考点看看对不
我以为你是客服。。。。。
确实。原图识别就有问题了。你只能提交工单反馈一下了。
你可以用我分享的百度网盘原图试试
https://pan.baidu.com/s/14y0ptmETaM5gLyWlwHW5lw
我想提供图片base64的编码,这边审核发送不了。图片上传的也不是原图,是被压缩后的结果
这个结果是这样的,我的图片上传到网页被压缩了,压缩前的图片识别不了
这是我那你上传到评论的图片测试的结果
通用OCR(高精度版): 耗时:00:00:01.1757824: { "log_id": 5442037564008187688, "direction": 1, "words_result_num": 29, "words_result": [ { "words": "美宜佳", "probability": { "variance": 0.061881, "average": 0.736606, "min": 0.432373 } }, { "words": " MEIYIJIA", "probability": { "variance": 0.0, "average": 0.858347, "min": 0.858347 } }, { "words": "店号:G743", "probability": { "variance": 0.004433, "average": 0.96752, "min": 0.805115 } }, { "words": "机号:G74301", "probability": { "variance": 0.000213, "average": 0.990521, "min": 0.959237 } }, { "words": "电话:0769-87792291", "probability": { "variance": 0.003882, "average": 0.974851, "min": 0.768595 } }, { "words": "收银:01店长", "probability": { "variance": 0.031838, "average": 0.929252, "min": 0.457246 } }, { "words": "时间:2018-10-1316:33:40", "probability": { "variance": 1.4E-05, "average": 0.997709, "min": 0.98636 } }, { "words": "代码", "probability": { "variance": 0.0, "average": 0.999895, "min": 0.999841 } }, { "words": "单价数量金额", "probability": { "variance": 0.0, "average": 0.999713, "min": 0.99938 } }, { "words": "121001709.50", "probability": { "variance": 1E-06, "average": 0.9993, "min": 0.997139 } }, { "words": "1.009.50", "probability": { "variance": 0.0, "average": 0.999522, "min": 0.998553 } }, { "words": "珠江桥牌广东南乳", "probability": { "variance": 4.1E-05, "average": 0.993939, "min": 0.981383 } }, { "words": "合计:1.00件", "probability": { "variance": 0.010106, "average": 0.956146, "min": 0.692657 } }, { "words": "9.50元", "probability": { "variance": 0.0, "average": 0.999415, "min": 0.998039 } }, { "words": "优惠:", "probability": { "variance": 3.1E-05, "average": 0.993183, "min": 0.986185 } }, { "words": "0.00元", "probability": { "variance": 0.0, "average": 0.999128, "min": 0.998197 } }, { "words": "应付:", "probability": { "variance": 0.000107, "average": 0.991475, "min": 0.976899 } }, { "words": "9.50元", "probability": { "variance": 0.0, "average": 0.99975, "min": 0.99905 } }, { "words": "通用扫码支付:", "probability": { "variance": 0.000188, "average": 0.98546, "min": 0.962083 } }, { "words": "9.50元", "probability": { "variance": 3E-06, "average": 0.998939, "min": 0.995423 } }, { "words": "找零:", "probability": { "variance": 0.000424, "average": 0.981005, "min": 0.95189 } }, { "words": "0.00元", "probability": { "variance": 1E-06, "average": 0.999445, "min": 0.998024 } }, { "words": "回", "probability": { "variance": 0.0, "average": 0.94668, "min": 0.94668 } }, { "words": "微信抖一扫", "probability": { "variance": 0.024988, "average": 0.829539, "min": 0.638805 } }, { "words": "券再买单", "probability": { "variance": 0.063318, "average": 0.78903, "min": 0.365216 } }, { "words": "优惠大大字", "probability": { "variance": 0.030289, "average": 0.883691, "min": 0.541861 } }, { "words": "美宜佳商城:yj.com.cn", "probability": { "variance": 0.000158, "average": 0.988892, "min": 0.95522 } }, { "words": "客服热线:400-887-1133", "probability": { "variance": 4.5E-05, "average": 0.995713, "min": 0.978105 } }, { "words": "公众号: myimyj520520", "probability": { "variance": 0.003303, "average": 0.975755, "min": 0.823772 } } ] }
我用api,http请求,也是一样的结果
有检测朝向
from aip import AipOcrdef get_file_content(filePath): with open(filePath, 'rb') as fp: return fp.read()
image = get_file_content('l14.jpg')
""" 调用通用文字识别(高精度版) """client = AipOcr(APP_ID, API_KEY, SECRET_KEY)# mresult=client.basicAccurate(image)# print(mresult)# print(client)
# """ 如果有可选参数 """options = {}options["detect_direction"] = "true"# options["probability"] = "true"## """ 带参数调用通用文字识别(高精度版) """mresult=client.basicAccurate(image, options)print(mresult)
记得点击那个箭头图标回复。我才能看到信息。你有给检测朝向那个参数吗?
可以把你的部分代码摘取出来的。尤其是发送的body参数那部分。
从大小可以看出 上传前的图片是137kb,上传后的是65.4kb
我把图片上传后再下载可以识别,但是上传前的哪张识别不了
用的是python sdk ,使用的是通用文字识别(高精度版)接口 返回的是:{'log_id': 6851876912516764331, 'direction': 0, 'words_result_num': 11, 'words_result': [{'words': '只只'}, {'words': '品'}, {'words': '正彩'}, {'words': 'N'}, {'words': '洲'}, {'words': '导23…'}, {'words': '寸'}, {'words': 'N'}, {'words': ','}, {'words': '二'}, {'words': '世≤'}]}
原图能发出来吗?还有你用的语言是什么?调用哪个接口。是API还是SDK。接口返回的信息又是什么呢?
这是调用API接口的问题?
您好,该问题已经解决,预计5个工作日内后端升级解决,前段无感知升级,您可以在10.26日重新测试线上接口。
该问题目前是badcase,已经优化,估计马上上线。
你可以识别同一张图片的同一参考点看看对不
我以为你是客服。。。。。
确实。原图识别就有问题了。你只能提交工单反馈一下了。
你可以用我分享的百度网盘原图试试
https://pan.baidu.com/s/14y0ptmETaM5gLyWlwHW5lw
我想提供图片base64的编码,这边审核发送不了。图片上传的也不是原图,是被压缩后的结果
这个结果是这样的,我的图片上传到网页被压缩了,压缩前的图片识别不了
这是我那你上传到评论的图片测试的结果
我用api,http请求,也是一样的结果
有检测朝向
from aip import AipOcr
def get_file_content(filePath):
with open(filePath, 'rb') as fp:
return fp.read()
image = get_file_content('l14.jpg')
""" 调用通用文字识别(高精度版) """
client = AipOcr(APP_ID, API_KEY, SECRET_KEY)
# mresult=client.basicAccurate(image)
# print(mresult)
# print(client)
# """ 如果有可选参数 """
options = {}
options["detect_direction"] = "true"
# options["probability"] = "true"
#
# """ 带参数调用通用文字识别(高精度版) """
mresult=client.basicAccurate(image, options)
print(mresult)
记得点击那个箭头图标回复。我才能看到信息。你有给检测朝向那个参数吗?
可以把你的部分代码摘取出来的。尤其是发送的body参数那部分。
从大小可以看出 上传前的图片是137kb,上传后的是65.4kb
我把图片上传后再下载可以识别,但是上传前的哪张识别不了
用的是python sdk ,使用的是通用文字识别(高精度版)接口 返回的是:{'log_id': 6851876912516764331, 'direction': 0, 'words_result_num': 11, 'words_result': [{'words': '只只'}, {'words': '品'}, {'words': '正彩'}, {'words': 'N'}, {'words': '洲'}, {'words': '导23…'}, {'words': '寸'}, {'words': 'N'}, {'words': ','}, {'words': '二'}, {'words': '世≤'}]}
原图能发出来吗?还有你用的语言是什么?调用哪个接口。是API还是SDK。接口返回的信息又是什么呢?