实现一个前后端的CS结构人脸识别小程序服务

2020-06-04 13:26:39 浏览数 (1)

一、实现方式:前端调用相机组件实现人脸在线采集,然后将人脸图片传到自建的服务端调用人脸识别-人脸检测与分析API将识别结果回调到小程序页面中。

二、实现流程

第一步:配置服务器域名

第二步:实现小程序的Demo

在小程序公共配置文件app.json中,添加页面生成参数

代码语言:json复制
"pages/camera/camera",

点击"编译"生成页面目录及页面

camera.wxml

代码语言:html复制
<!-- camera.wxml -->
<camera device-position="front" flash="off" binderror="error" style="width: 100%; height: 300px;"></camera>
<button type="primary" bindtap="takePhoto">拍照</button>
<view wx:if="{{ FaceInfos['0']['FaceAttributesInfo']['Gender'] > 50}}">性别:男</view>
<view wx:if="{{ FaceInfos['0']['FaceAttributesInfo']['Gender'] < 50}}">性别:女</view>
<view>年龄:{{ FaceInfos['0']['FaceAttributesInfo']['Age'] }}</view>
<view wx:if="{{ FaceInfos['0']['FaceAttributesInfo']['Expression'] == 0 }}">表情:正常</view>
<view wx:if="{{ FaceInfos['0']['FaceAttributesInfo']['Expression'] < 50 }}">表情:微笑</view>
<view wx:if="{{ FaceInfos['0']['FaceAttributesInfo']['Expression'] > 50 }}">表情: 大笑</view>
<view wx:if="{{ FaceInfos['0']['FaceAttributesInfo']['Beauty'] == 0 }}">魅力值:一般</view>
<view wx:if="{{ FaceInfos['0']['FaceAttributesInfo']['Beauty'] < 50 }}">魅力值:有点迷人</view>
<view wx:if="{{ FaceInfos['0']['FaceAttributesInfo']['Beauty'] > 50 }}">魅力值:偶像级</view>
<view>请求的图片宽度:{{ ImageWidth }}</view>
<view>请求的图片高度:{{ ImageHeight }}</view>
<view>人脸框左上角横坐标:{{ FaceInfos['0']['X'] }}</view>
<view>人脸框左上角纵坐标:{{ FaceInfos['0']['Y'] }}</view>
<view>人脸框宽度:{{ FaceInfos['0']['Width'] }}</view>
<view>人脸框高度:{{ FaceInfos['0']['Height'] }}</view>
<image mode="widthFix" src="{{src}}"></image>

使用的组件:

camera

button

image

使用的视图容器:

view

使用的XML语法:

wx:if条件渲染

双大括号数据绑定

使用的视图层:

bindtap事件绑定

WXSS样式学习

camera.json

代码语言:json复制
{
  "navigationBarTitleText": "人脸识别在线测试",
  "backgroundColor": "#eeeeee"
}

全局配置

camera.js

代码语言:javascript复制
// camera.js
Page({
  takePhoto() {
    var that=this;
    const ctx = wx.createCameraContext()
    ctx.takePhoto({
      quality: 'high',
      success: (res) => {
        this.setData({
          src: res.tempImagePath
        })
        wx.request({
          url: 'https://tencentcloud.cdhwdl.com:8000', //仅为示例,并非真实的接口地址
          method:'post',
          data: {
            x: "data:image/jpg;base64,"   wx.getFileSystemManager().readFileSync(res.tempImagePath, 'base64')
          },
          header: {
            'content-type': 'application/json' // 默认值
          },
          success (res) {
            that.setData({
              ImageWidth: res.data.Result.ImageWidth "px",
              ImageHeight: res.data.Result.ImageHeight "px",
              FaceInfos: res.data.Result.FaceInfos,
            })
            console.log(res.data)

          }
        })
      }
    })
  },
  error(e) {
    console.log(e.detail)
  }
})

使用到的知识点: Page 构造器

创建 camera 上下文 CameraContext 对象

HTTPS 网络请求

setData

注意:如果自定义函数中嵌套了wx等对象函数,数据传递应该先声明"var that=this",然后再嵌套函数,如wx.request中使用"that.setData"来传递数据

后端数据结构

第三步:搭建nodejs服务端

任意安装一款Linux发行版系统(安装过程略)

代码语言:javascript复制
[root@zhang .nvm]# cat /etc/redhat-release 
CentOS release 6.9 (Final)

安装2.0版本以上的git客户端,如果你的系统是Centos发行版的,可以参考下面的安装演示;如果是其他发行版,可以参考git官网指引,通过简单的命令即可安装

非Centos发行版系统安装方式参考Git官方文档下载指引

Centos发行版系统(这里是Centos6.9)安装流程如下:

安装Git依赖包:

检查是否安装"Development Tools"软件组,若未安装则执行安装命令

代码语言:javascript复制
[root@zhang tmp]# yum grouplist | grep "Development Tools"
[root@zhang tmp]# 
[root@zhang yum.repos.d]# yum  groupinstall "Development Tools" -y
[root@zhang yum.repos.d]# yum grouplist | grep "Development tools"
   Development tools

安装其他软件包(如果已安装了会提示已安装)

代码语言:javascript复制
yum install zlib-devel -y
yum install perl-ExtUtils-MakeMaker -y
yum install asciidoc -y 
yum install xmlto -y
yum install openssl-devel -y 
yum install gcc -y 
yum install curl-devel -y 
yum install expat-devel -y 
yum install gettext-devel -y

卸载现有Git

代码语言:javascript复制
[root@zhang git-2.0.5]# yum remove git -y

下载2.0版本的Git客户端,如果下载慢,可以用网速较好的机器下载后再上传到服务器中,下载后解压

代码语言:javascript复制
[root@zhang tmp]# wget https://www.kernel.org/pub/software/scm/git/git-2.0.5.tar.gz
代码语言:javascript复制
[root@zhang tmp]# ls -lh | grep git
drwxrwxr-x 19 root  root   12K Dec 19  2014 git-2.0.5
-rw-r--r--  1 root  root  4.7M Dec 19  2014 git-2.0.5.tar.gz

进入解压目录,三步编译安装法安装

软件配置与检查

代码语言:javascript复制
[root@zhang git-2.0.5]# ./configure --prefix=/usr/local/git

编译成二进制文件

代码语言:javascript复制
[root@zhang git-2.0.5]# make

安装编译后的文件到指定目录

代码语言:javascript复制
[root@zhang git-2.0.5]# make install

将Git的运行程序路径配置到全局环境变量中(路径为"/usr/local/git/bin")

代码语言:javascript复制
[root@zhang git-2.0.5]# vi /etc/profile
代码语言:javascript复制
[root@zhang git-2.0.5]# cat /etc/profile | grep "export PATH="
export PATH=/usr/local/nginx/sbin:/usr/local/php/bin:/usr/local/mysql/bin:$PATH:/usr/local/git/bin

使得修改生效

代码语言:javascript复制
[root@zhang git-2.0.5]# source /etc/profile
[root@zhang git-2.0.5]# 

查看git版本号

代码语言:javascript复制
[root@zhang git-2.0.5]# git --version
git version 2.0.5

安装nvm

参考官方文档:https://github.com/nvm-sh/nvm/blob/master/README.md

代码语言:javascript复制
[root@zhang git-2.0.5]# curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.35.3/install.sh | bash
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 13527  100 13527    0     0   5385      0  0:00:02  0:00:02 --:--:-- 12131
=> Downloading nvm from git to '/root/.nvm'
=> Cloning into '/root/.nvm'...
remote: Enumerating objects: 290, done.
remote: Counting objects: 100% (290/290), done.
remote: Compressing objects: 100% (257/257), done.
remote: Total 290 (delta 35), reused 97 (delta 20), pack-reused 0
Receiving objects: 100% (290/290), 163.27 KiB | 8.00 KiB/s, done.
Resolving deltas: 100% (35/35), done.
Checking connectivity... done.
=> Compressing and cleaning up git repository

=> Appending nvm source string to /root/.bashrc
=> Appending bash_completion source string to /root/.bashrc
=> Close and reopen your terminal to start using nvm or run the following to use it now:

export NVM_DIR="$HOME/.nvm"
[ -s "$NVM_DIR/nvm.sh" ] && . "$NVM_DIR/nvm.sh"  # This loads nvm
[ -s "$NVM_DIR/bash_completion" ] && . "$NVM_DIR/bash_completion"  # This loads nvm bash_completion

在当前用户的环境变量配置文件"~/.bash_profile"或者全局环境变量配置文件"/etc/profile"中加入如下内容

export NVM_DIR="$([ -z "${XDG_CONFIG_HOME-}" ] && printf %s "${HOME}/.nvm" || printf %s "${XDG_CONFIG_HOME}/nvm")" [ -s "$NVM_DIR/nvm.sh" ] && . "$NVM_DIR/nvm.sh" # This loads nvm

代码语言:javascript复制
[root@zhang ~]# vi ~/.bash_profile
[root@zhang ~]# tail -2f ~/.bash_profile 
export NVM_DIR="$([ -z "${XDG_CONFIG_HOME-}" ] && printf %s "${HOME}/.nvm" || printf %s "${XDG_CONFIG_HOME}/nvm")"
[ -s "$NVM_DIR/nvm.sh" ] && . "$NVM_DIR/nvm.sh"

重载环境变量

代码语言:javascript复制
source ~/.bash_profile

测试nvm是否安装成功

代码语言:javascript复制
[root@zhang ~]# nvm --version
0.35.3
[root@zhang ~]# 

安装Node.js 7.10.1 版本及以上

代码语言:javascript复制
[root@zhang iai]# nvm install v10.6.0
Downloading and installing node v10.6.0...
Downloading https://nodejs.org/dist/v10.6.0/node-v10.6.0-linux-x64.tar.xz...
######################################################################## 100.0%
Computing checksum with sha256sum
Checksums matched!
Now using node v10.6.0 (npm v6.1.0)
[root@zhang iai]# node -v
v10.6.0

创建nodejs web项目仓库

代码语言:shell复制
[root@zhang data]# mkdir -p /data/nodejs

创建人脸识别项目

代码语言:shell复制
[root@zhang data]# mkdir /data/nodejs/iai
[root@zhang data]# cd /data/nodejs/iai
[root@zhang data]# mkdir picture
[root@zhang data]# chmod -R 775 picture

安装tencentcloud-sdk-nodejs并创建一个项目入口文件app.js

代码语言:javascript复制
[root@zhang iai]# npm install tencentcloud-sdk-nodejs --save
[root@zhang iai]# ls -lh
total 28K
-rw-r--r--  1 root root  249 Apr 14 15:18 app.js
drwxr-xr-x 51 root root 4.0K Apr 14 15:39 node_modules
-rw-r--r--  1 root root 2.7K Apr 14 09:53 npm-debug.log
-rw-r--r--  1 root root  13K Apr 14 15:39 package-lock.json

通过API 3.0 Explorer生成tencentcloud-sdk-nodejs下人脸检测与分析API的调用Demo

https://console.cloud.tencent.com/api/explorer?Product=iai&Version=2018-03-01&Action=DetectFace&SignVersion=

在实现Web功能之前,我们需要知道小程序的服务端只允许HTTPS协议的地址,所以我们应该通过nodejs的HTTPS模块来实现一个加密的Web服务,具体流程如下:

  • 通过一个已经实名认证的腾讯云账号在控制台进入“SSL证书”控制台,点击【申请免费证书】为你的小程序服务端域名免费申请一个SSL加密证书

  • 申请成功后下载证书文件压缩包
  • 解压缩后进入到Nginx目录下

  • 在Linux服务端nodejs的项目目录下创建certificate目录并配置权限755
代码语言:shell复制
[root@zhang iai]# ls -lh
total 32K
-rwxrwxrwx  1 root root 2.2K Apr 15 10:17 app.js
drwxrwxrwx 51 root root 4.0K Apr 14 15:39 node_modules
-rw-------  1 root root    0 Apr 15 09:50 nohup.out
-rwxrwxrwx  1 root root 2.7K Apr 14 09:53 npm-debug.log
-rwxrwxrwx  1 root root  13K Apr 14 15:39 package-lock.json
drwxrwxrwx  2 root root 4.0K Apr 15 10:31 picture
[root@zhang iai]# mkdir certificate
[root@zhang iai]# chmod 775 certificate/
  • 上传Nginx目录下的两个证书文件到服务端的certificate目录下并重名为"server.key"、"server.crt"
代码语言:shell复制
[root@zhang certificate]# ls -lh
total 8.0K
-rw-r--r-- 1 root root 3.7K Apr 15 10:48 1_tencentcloud.cdhwdl.com_bundle.crt
-rw-r--r-- 1 root root 1.7K Apr 15 10:48 2_tencentcloud.cdhwdl.com.key
[root@zhang certificate]# mv 1_tencentcloud.cdhwdl.com_bundle.crt server.crt
[root@zhang certificate]# mv 2_tencentcloud.cdhwdl.com.key server.key
[root@zhang certificate]# ls -lh
total 8.0K
-rwxr-xr-x 1 root root 3.7K Apr 15 10:48 server.crt
-rwxr-xr-x 1 root root 1.7K Apr 15 10:48 server.key

编辑项目入口文件app.js,通过nodejs的https模块创建一个web服务器并调用上述Demo

代码语言:javascript复制
const https = require('https');
const fs = require('fs');
const path = require('path');

const privateKey  = fs.readFileSync(path.join(__dirname, './certificate/server.key'), 'utf8');
const certificate = fs.readFileSync(path.join(__dirname, './certificate/server.crt'), 'utf8');
const credentials = {key: privateKey, cert: certificate};
const httpsServer = https.createServer(credentials,function(req, res){
    let body = [];
    req.on('data', (chunk) => {
        body.push(chunk);
    }).on('end', () => {
        body = Buffer.concat(body).toString();
        res.statusCode = 200;
        res.setHeader('Content-Type', 'text/plain');
        var json_ob = JSON.parse(body.trim());
        base64 = json_ob.x;
        var base64Data = base64.replace(/^data:image/w ;base64,/, "");
        var dataBuffer = Buffer.from(base64Data, 'base64');
        var t1 = new Date().getTime();
        var name = t1 ".jpg";
        var json = { Result: ""};
        fs.writeFile("./picture/" name, dataBuffer, { 'flag': 'a' }, function(err) {
            if(err){
                json.Result=err;
                res.end(JSON.stringify(json));
            }
        });
        const tencentcloud = require("./node_modules/tencentcloud-sdk-nodejs");


        const IaiClient = tencentcloud.iai.v20180301.Client;
        const models = tencentcloud.iai.v20180301.Models;

        const Credential = tencentcloud.common.Credential;
        const ClientProfile = tencentcloud.common.ClientProfile;
        const HttpProfile = tencentcloud.common.HttpProfile;

        let cred = new Credential("", "");
        let httpProfile = new HttpProfile();
        httpProfile.endpoint = "iai.tencentcloudapi.com";
        let clientProfile = new ClientProfile();
        clientProfile.httpProfile = httpProfile;
        let client = new IaiClient(cred, "ap-guangzhou", clientProfile);

        let req = new models.DetectFaceRequest();
        var params = {Image: base64Data, NeedFaceAttributes: 1}
        var params = JSON.stringify(params)
        req.from_json_string(params);


        client.DetectFace(req, function(errMsg, response) {

            if (errMsg) {
                json.Result=errMsg;
                res.end(JSON.stringify(json));
            }
            json.Result=JSON.parse(response.to_json_string());
            res.end(JSON.stringify(json));
        });
    });

});


const SSLPORT = 8000;
httpsServer.listen(SSLPORT, '0.0.0.0', () => {});

服务端后台启用这个web服务

代码语言:shell复制
[root@zhang iai]# nohup node app.js &
[1] 4446
nohup: ignoring input and appending output to `nohup.out'

我们先通过postman来在线调试下

测试OK,可以将web服务地址对接到小程序人脸识别页面中了

测试效果:

0 人点赞