OCR 百度AI 文字识别

10.00 触摸精灵调用百度AI实现文字识别

发布于 2019-06-20 21:29:18

1.注册一个百度智能云账号 https://cloud.baidu.com/

2.登录进入控制台选择人工智能下面的文字识别

3.创建一个应用,名字类型应用描述随便写.接口选择不用管默认就可以.

4.返回应用列表后记住 API Key 和 Secret Key 这两个的内容.对应下面脚本内的AK 和 SK

5.点击概览可以看到各种识别可以免费用几次,普通的通用文字识别一天是5万次的免费,并发数是10,正常使用应该会够,不够怎么办?多注册几个账号切换用呗

6.文字识别文档:https://cloud.baidu.com/doc/OCR/s/zjwvxzr65

7.注意!!!!使用此函数必须使用373以上版本!!!!!

BaiduAi = {
    Init = (function()
        local AK = '这里写应用的API Key';
        local SK = '这里写应用的Secret Key';
        local AccessToken = httpPost('https://aip.baidubce.com/oauth/2.0/token',{["grant_type"]="client_credentials",["client_id"]=AK,["client_secret"]=SK})
        AccessToken = jsonDecode(AccessToken);
        return AccessToken.access_token;
    end),
    Ocr = (function(ImagePath,AT,Api,Ext)
        local file = io.open(ImagePath, 'rb');
        local data = file:read("*all");
        file:close();
        local image = urlEncode(Base64.Encode(data));
        Ext = Ext or {};
        Ext["image"]=image;
        local OcrInfo = httpPost(string.format('https://aip.baidubce.com/rest/2.0/ocr/v1/%s?access_token=%s',Api,AT),Ext);
        return OcrInfo;
    end)

}
Base64 = {
    Init = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/',
    Encode = (function(data)
            local b = Base64.Init;
            return ((data:gsub('.', function(x)
                local r,b='',x:byte()
                for i=8,1,-1 do r=r..(b%2^i-b%2^(i-1)>0 and '1' or '0') end
                return r;
            end)..'0000'):gsub('%d%d%d?%d?%d?%d?', function(x)
                if (#x < 6) then return '' end
                local c=0
                for i=1,6 do c=c+(x:sub(i,i)=='1' and 2^(6-i) or 0) end
                return b:sub(c+1,c+1)
            end)..({ '', '==', '=' })[#data%3+1])
        end),
    Decode = (function(data)
        local b = Base64.Init;
        data = string.gsub(data, '[^'..b..'=]', '')
        return (data:gsub('.', function(x)
            if (x == '=') then return '' end
            local r,f='',(b:find(x)-1)
            for i=6,1,-1 do r=r..(f%2^i-f%2^(i-1)>0 and '1' or '0') end
            return r;
        end):gsub('%d%d%d?%d?%d?%d?%d?%d?', function(x)
            if (#x ~= 8) then return '' end
            local c=0
            for i=1,8 do c=c+(x:sub(i,i)=='1' and 2^(8-i) or 0) end
            return string.char(c)
        end))
    end)
}
function urlEncode(s)
     s = string.gsub(s, "([^%w%.%- ])", function(c) return string.format("%%%02X", string.byte(c)) end)
    return string.gsub(s, " ", "+")
end
function httpPost(path,data)
    --local https = require("ssl.https") --5.0以下版本加载ssl才能访问https链接
    local https = require("socket.http") --5.0版本以上触摸加载http即可
    local ltn12 = require("ltn12")
    local request_body = '';
        if type(data)=='table' then
            for k,v in pairs(data) do
                request_body = request_body..k..'='..v..'&'
            end
        else
            request_body = data
        end
    local response_body = {}
    local res, code, response_headers = https.request{
        url = path,
        method = "POST",
        headers =
          {
              ["Content-Type"] = "application/x-www-form-urlencoded";
              ["Content-Length"] = string.len(request_body);
         },
          source = ltn12.source.string(request_body),
          sink = ltn12.sink.table(response_body),
    }
    if type(response_body) == "table" then
      if table.concat(response_body) ~= "error" then
          return table.concat(response_body);
      else
          return -1;
      end
    else
      logDebug("Not a table:", type(response_body))
      return -1;
    end
end
function main()
    mSleep(2000)
    --获取AccessToken,这个AccessToken有效期为30天,没必要获取很多次,可以获取一次保存下来
    AccessToken = BaiduAi.Init()
    logDebug(AccessToken)
    ImagePath = '/var/touchelf/res/1.jpg'
    --截图,也可以使用区域截图,注意图片的要求
    --图像数据，base64编码后进行urlencode，要求base64编码和urlencode后大小不超过4M，最短边至少15px，最长边最大4096px
    snapshotScreen(ImagePath);
    --通用文字识别
    Ocr = BaiduAi.Ocr(ImagePath,AccessToken,'general_basic')
    --返回的是个json数据
    logDebug(Ocr)
    --通用文字识别（高精度含位置版）
    Ocr = BaiduAi.Ocr(ImagePath,AccessToken,'accurate')
    --返回的是个json数据
    logDebug(Ocr)
    --通用文字识别自定义参数写法
    Ocr = BaiduAi.Ocr(ImagePath,AccessToken,'general_basic',{["vertexes_location"]='true',["probability"]='true'})
    --返回的是个json数据
    logDebug(Ocr)
    --其他类型看文档修改BaiduAi.Ocr第三个参数即可.
end

通用文字识别返回示例

[2019-06-20 21:07:46] {"log_id": 745660733214676500, "words_result_num": 16, "words_result": [{"words": "无SM卡令"}, {"words": "下午9:07"}, {"words": "触摸精灵"}, {"words": "信息"}, {"words": "关于"}, {"words": "帮助"}, {"words": "系统信息"}, {"words": "查询已安装应用包名"}, {"words": "查看日志"}, {"words": "清空日志"}, {"words": "清空GPS伪装信息"}, {"words": ">>>"}, {"words": "重启设备"}, {"words": "三"}, {"words": "列表"}, {"words": "设置"}]}
通用文字识别（高精度含位置版）返回示例
[2019-06-20 21:10:17] {"log_id": 7008446590512061492, "words_result_num": 20, "words_result": [{"location": {"width": 142, "top": 6, "left": 11, "height": 27}, "words": "无SM卡令"}, {"location": {"width": 97, "top": 7, "left": 274, "height": 26}, "words": "下午9:10"}, {"location": {"width": 136, "top": 65, "left": 251, "height": 37}, "words": "触摸精灵"}, {"location": {"width": 34, "top": 264, "left": 30, "height": 123}, "words": "关帮"}, {"location": {"width": 19, "top": 268, "left": 590, "height": 28}, "words": ">"}, {"location": {"width": 31, "top": 352, "left": 65, "height": 36}, "words": "助"}, {"location": {"width": 18, "top": 357, "left": 591, "height": 27}, "words": ">"}, {"location": {"width": 105, "top": 482, "left": 29, "height": 29}, "words": "系统信息"}, {"location": {"width": 308, "top": 551, "left": 30, "height": 37}, "words": "查询已安装应用包名"}, {"location": {"width": 19, "top": 555, "left": 591, "height": 28}, "words": ">"}, {"location": {"width": 136, "top": 638, "left": 30, "height": 37}, "words": "查看日志"}, {"location": {"width": 18, "top": 643, "left": 591, "height": 28}, "words": ">"}, {"location": {"width": 136, "top": 727, "left": 30, "height": 36}, "words": "清空日志"}, {"location": {"width": 18, "top": 731, "left": 591, "height": 28}, "words": ">"}, {"location": {"width": 271, "top": 814, "left": 30, "height": 38}, "words": "清空GPS伪装信息"}, {"location": {"width": 18, "top": 819, "left": 591, "height": 29}, "words": ">"}, {"location": {"width": 136, "top": 903, "left": 30, "height": 36}, "words": "重启设备"}, {"location": {"width": 18, "top": 907, "left": 591, "height": 28}, "words": ">"}, {"location": {"width": 41, "top": 1110, "left": 137, "height": 24}, "words": "列表"}, {"location": {"width": 39, "top": 1111, "left": 458, "height": 22}, "words": "设置"}]}

使用触摸自带httpPost

BaiduAi = {
    Init = (function()
        local AK = '这里写应用的API Key';
        local SK = '这里写应用的Secret Key';
        local AccessToken = httpPost('https://aip.baidubce.com/oauth/2.0/token',{},{["grant_type"]="client_credentials",["client_id"]=AK,["client_secret"]=SK},10)
        AccessToken = jsonDecode(AccessToken);
        return AccessToken.access_token;
    end),
    Ocr = (function(ImagePath,AT,Api,Ext)
        local file = io.open(ImagePath, 'rb');
        local data = file:read("*all");
        file:close();
        local image = urlEncode(Base64.Encode(data));
        Ext = Ext or {};
        Ext["image"]=image;
        local OcrInfo = httpPost(string.format('https://aip.baidubce.com/rest/2.0/ocr/v1/%s?access_token=%s',Api,AT),{},Ext,10);
        return OcrInfo;
    end)

}
Base64 = {
    Init = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/',
    Encode = (function(data)
            local b = Base64.Init;
            return ((data:gsub('.', function(x)
                local r,b='',x:byte()
                for i=8,1,-1 do r=r..(b%2^i-b%2^(i-1)>0 and '1' or '0') end
                return r;
            end)..'0000'):gsub('%d%d%d?%d?%d?%d?', function(x)
                if (#x < 6) then return '' end
                local c=0
                for i=1,6 do c=c+(x:sub(i,i)=='1' and 2^(6-i) or 0) end
                return b:sub(c+1,c+1)
            end)..({ '', '==', '=' })[#data%3+1])
        end),
    Decode = (function(data)
        local b = Base64.Init;
        data = string.gsub(data, '[^'..b..'=]', '')
        return (data:gsub('.', function(x)
            if (x == '=') then return '' end
            local r,f='',(b:find(x)-1)
            for i=6,1,-1 do r=r..(f%2^i-f%2^(i-1)>0 and '1' or '0') end
            return r;
        end):gsub('%d%d%d?%d?%d?%d?%d?%d?', function(x)
            if (#x ~= 8) then return '' end
            local c=0
            for i=1,8 do c=c+(x:sub(i,i)=='1' and 2^(8-i) or 0) end
            return string.char(c)
        end))
    end)
}
function urlEncode(s)
     s = string.gsub(s, "([^%w%.%- ])", function(c) return string.format("%%%02X", string.byte(c)) end)
    return string.gsub(s, " ", "+")
end
function main()
    mSleep(2000)
    --获取AccessToken,这个AccessToken有效期为30天,没必要获取很多次,可以获取一次保存下来
    local AccessToken = BaiduAi.Init()
    logDebug(AccessToken)
    ImagePath = '/var/touchelf/1.jpg'
    --截图,也可以使用区域截图,注意图片的要求
    --图像数据，base64编码后进行urlencode，要求base64编码和urlencode后大小不超过4M，最短边至少15px，最长边最大4096px
    snapshotScreen(ImagePath);
    --通用文字识别
    Ocr = BaiduAi.Ocr(ImagePath,AccessToken,'general_basic')
    --返回的是个json数据
    logDebug(Ocr)
end

18 条评论

fstone 2019-06-22

大神我想用数字识别的话应该怎么改？

Evan 2019-06-22

@fstone 不用改.直接识别就是了.注意图片大小的要求

fstone 2019-06-22

不行啊识别不到数字啊！他有个专门的数字识别，能教我一下改改吗？

fstone 2019-06-22

@evan 不行啊识别不到数字啊！他有个专门的数字识别，能教我一下改改吗？

Evan 2019-06-23

@fstone Ocr = BaiduAi.Ocr(ImagePath,AccessToken,'numbers')

cxb746 2019-11-06

local https = require("ssl.https")在这行提示出错，貌似找不到插件包

cxb746 2019-11-06

@evan local https = require("ssl.https")在这行提示出错，貌似找不到插件包

Evan 2019-11-07

@cxb746 要求触摸最低版本373以上

monv22 2020-01-29

@evan
Ocr = BaiduAi.Ocr（ImagePath，AccessToken，'数字'）
{"error_code":3,"error_msg":"Unsupported openapi method"}
这段代码运行是错误的,
请问调用对比数字或者字母的简单函数是什么,本人小白请见谅!
我的理解是
AccessToken = BaiduAi.Init()

logDebug(AccessToken)
ImagePath = '/var/touchelf/res/1.jpg'
snapshotScreen(ImagePath);
Ocr = BaiduAi.Ocr(ImagePath,AccessToken,"1")--图片里数字为1
    if Ocr == "" then
    ts("没有识别到结果")
    elseifOcr == 1 then
    ts("找到")
end
类似这样的要怎么写,这段代码是错误的
非常需要这个功能,求指点