当前位置：网站首页>基于PaddleOCR开发uni-app离线身份证识别插件
基于PaddleOCR开发uni-app离线身份证识别插件
2022-07-25 14:13:00 【番茄小能手】
目录
1、在android sudio创建一个项目，并创建一个Module
2、将ocr插件和uniapp-v8-release.aar引入libs目录下
目的
之前做了基于PaddleOCR的DBNet多分类文本检测网络之身份证识别，可以得到一个不超过3M大小的模型，通过了解PaddleOCR的端侧部署，我们也可以将身份证检测模型移植到手机中，做成一款uni-app手机端离线身份证识别的应用。
准备工作
为了不占用过多篇幅，这里不讲解Android studio如何下载和SDK配置，请自行查找资料。
1、HbuilderX
2、Android Studio 并配置 NDK
请根据 Android Studio 用户指南中的安装及配置 NDK 和 CMake 内容，预先配置好 NDK 。
3、下载基于PaddleOCR的离线身份证识别插件
插件包结构：
开始
1、在android sudio创建一个项目，并创建一个Module
2、将ocr插件和uniapp-v8-release.aar引入libs目录下
3、修改module包下的build.gradle
apply plugin: 'com.android.library'
android {
compileSdkVersion 32
defaultConfig {
minSdkVersion 21
targetSdkVersion 32
versionCode 1
versionName "1.0"
testInstrumentationRunner "androidx.test.runner.AndroidJUnitRunner"
consumerProguardFiles "consumer-rules.pro"
}
buildTypes {
release {
minifyEnabled false
proguardFiles getDefaultProguardFile('proguard-android-optimize.txt'), 'proguard-rules.pro'
}
}
sourceSets {
main {
jniLibs.srcDirs = ['libs']
}
}
repositories {
flatDir {
dirs('libs')
}
}
compileOptions {
sourceCompatibility JavaVersion.VERSION_1_8
targetCompatibility JavaVersion.VERSION_1_8
}
}
dependencies {
implementation fileTree(dir: "libs", include: ["*.jar"])
implementation 'androidx.appcompat:appcompat:1.2.0'
testImplementation 'junit:junit:4.12'
androidTestImplementation 'androidx.test.ext:junit:1.1.3'
androidTestImplementation 'androidx.test.espresso:espresso-core:3.4.0'
compileOnly(name: 'uniapp-v8-release', ext: 'aar')
compileOnly 'com.alibaba:fastjson:1.1.46.android'
implementation(name: 'ocr', ext: 'aar')
}4、添加OCRModule.java文件
public class OCRModule extends UniModule {
private final String TYPE_IDCARD = "idcard";
public OCRPredictor ocrPredictor = null;
public UniJSCallback callback = null;
@UniJSMethod(uiThread = true)
public void ocrAsyncFunc(JSONObject options, final UniJSCallback callback) {
Log.d("OCRModule", "ocrAsyncFunc--" + options);
this.callback = callback;
ocrPredictor = OCRPredictor.getInstance(mUniSDKInstance.getContext());
if (TYPE_IDCARD.equals(options.getString("type"))) {
idcard();
} else {
if (options.containsKey("filePath")) {
String filePath = options.getString("filePath");
if (!TextUtils.isEmpty(filePath) && filePath.contains("file://")) {
filePath = filePath.replace("file://", "");
}
ocrPredictor.predictor(filePath, new OnImagePredictorListener() {
@Override
public void success(String result, ArrayList<OCRResultModel> ocrResultModelList, Bitmap bitmap) {
JSONArray jsonArray = new JSONArray();
for (OCRResultModel ocrResultModel : ocrResultModelList) {
JSONObject jsonObject = new JSONObject();
if (!TextUtils.isEmpty(ocrResultModel.getLabel())) {
jsonObject.put("words", ocrResultModel.getLabel());
JSONArray objects = new JSONArray();
for (Point point : ocrResultModel.getPoints()) {
JSONArray points = new JSONArray();
points.add(point.x);
points.add(point.y);
objects.add(points);
}
jsonObject.put("location", objects);
jsonObject.put("score", ocrResultModel.getConfidence());
jsonArray.add(jsonObject);
}
}
String jsonString = jsonArray.toJSONString();
Log.d("OCRModule", "ocrResult--" + result);
Log.d("OCRModule", "ocrResult--" + jsonString);
callback.invoke(jsonString);
}
});
} else if (options.containsKey("base64")) {
String base64 = options.getString("base64");
ocrPredictor.predictorBase64(base64, new OnImagePredictorListener() {
@Override
public void success(String result, ArrayList<OCRResultModel> ocrResultModelList, Bitmap bitmap) {
JSONArray jsonArray = new JSONArray();
for (OCRResultModel ocrResultModel : ocrResultModelList) {
JSONObject jsonObject = new JSONObject();
if (!TextUtils.isEmpty(ocrResultModel.getLabel())) {
jsonObject.put("words", ocrResultModel.getLabel());
JSONArray objects = new JSONArray();
for (Point point : ocrResultModel.getPoints()) {
JSONArray points = new JSONArray();
points.add(point.x);
points.add(point.y);
objects.add(points);
}
jsonObject.put("location", objects);
jsonObject.put("score", ocrResultModel.getConfidence());
jsonArray.add(jsonObject);
}
}
String jsonString = jsonArray.toJSONString();
Log.d("OCRModule", "ocrResult--" + result);
Log.d("OCRModule", "ocrResult--" + jsonString);
callback.invoke(jsonString);
}
});
} else {
this.callback.invoke("");
}
}
}
private static final int REQUEST_CODE_PICK_IMAGE = 200;
private static final int REQUEST_CODE_PICK_IMAGE_FRONT = 201;
private static final int REQUEST_CODE_PICK_IMAGE_BACK = 202;
private static final int REQUEST_CODE_CAMERA = 102;
private void idcard() {
Intent intent = new Intent(mUniSDKInstance.getContext(), CameraActivity.class);
intent.putExtra(CameraActivity.KEY_CONTENT_TYPE, CameraActivity.CONTENT_TYPE_ID_CARD);
((Activity) mUniSDKInstance.getContext()).startActivityForResult(intent, REQUEST_CODE_PICK_IMAGE);
}
@Override
public void onActivityResult(int requestCode, int resultCode, Intent data) {
if (REQUEST_CODE_PICK_IMAGE == requestCode) {
if (data != null) {
String result = data.getStringExtra(CameraActivity.KEY_CONTENT_RESULT);
Log.d("OCRModule", "ocrResult--" + result);
this.callback.invoke(result);
} else {
this.callback.invoke("");
}
} else {
super.onActivityResult(requestCode, resultCode, data);
}
}
}5、将module打包aar包
6、打开Hbuilder新建一个uni-app项目
将插件包和我们打的aar包如下放置：
注：这里ocr_module_support-release.aar改名为TomatoOCR.aar，改不改名无所谓。
7、修改package.json中的配置
注意：name名是我们的插件名，class是我们编写的识别入口的类
"plugins": [
{
"type": "module",
"name": "YY-TomatoOCR",
"class": "com.tomato.ocr.OCRModule"
}
],
{
"name": "YY-TomatoOCR",
"id": "YY-TomatoOCR",
"version": "1.0.3",
"description": "离线版文字识别",
"_dp_type": "nativeplugin",
"_dp_nativeplugin": {
"android": {
"plugins": [
{
"type": "module",
"name": "YY-TomatoOCR",
"class": "com.tomato.ocr.OCRModule"
}
],
"hooksClass": "",
"integrateType": "aar",
"dependencies": [
],
"excludeDependencies": [],
"compileOptions": {
"sourceCompatibility": "1.8",
"targetCompatibility": "1.8"
},
"abis": [
"armeabi-v7a",
"arm64-v8a"
],
"minSdkVersion": "21",
"useAndroidX": true,
"permissions": [
"android.permission.CAMERA",
"android.permission.READ_EXTERNAL_STORAGE",
"android.permission.WRITE_EXTERNAL_STORAGE"
]
}
}
}8、在mainifest.json配置中引入我们的原生插件
9、编写uni-app页面代码
<template>
<div>
<button type="primary" @click="takePhoto">拍照OCR识别</button>
<button type="primary" @click="idcard">身份证识别</button>
</div>
</template>
<script>
var ocrModule = uni.requireNativePlugin("YY-TomatoOCR")
// var ocrModule = uni.requireNativePlugin("OCRModule")
const modal = uni.requireNativePlugin('modal');
export default {
onLoad() {},
methods: {
takePhoto() {
uni.chooseImage({
count: 6, //默认9
sizeType: ['original', 'compressed'], //可以指定是原图还是压缩图，默认二者都有
sourceType: ['camera'], //从相册选择
success: function(res) {
console.log(JSON.stringify(res.tempFilePaths));
uni.getImageInfo({
src: res.tempFilePaths[0],
success: function(image) {
// 调用异步方法
ocrModule.ocrAsyncFunc({
'filePath': image.path
},
(ret) => {
modal.toast({
message: ret,
duration: 30
});
})
}
});
}
});
},
idcard() {
ocrModule.ocrAsyncFunc({
'type': 'idcard'
},
(ret) => {
modal.toast({
message: ret,
duration: 30
});
})
}
}
}
</script>10、制作自定义基座打包并运行到手机上
完毕！！！
总结
以上就是制作uni-app离线文字识别和身份证识别全流程步骤，当然如果你不会Android开发也没关系，在uni-app的插件市场中已经发布了本插件。
边栏推荐
- [directory blasting tool] information collection stage: robots.txt, Yujian, dirsearch, dirb, gobuster
- Four methods of importing CSV text files into Excel
- 力扣（LeetCode）205. 同构字符串（2022.07.24）
- Digital Twins - cognition
- CTS测试介绍(面试怎么介绍接口测试)
- 【目录爆破工具】信息收集阶段：robots.txt、御剑、dirsearch、Dirb、Gobuster
- PT100 temperature measurement circuit diagram (AD590 typical temperature measurement circuit)
- Hyperautomation for the enhancement of automation in industries
- AI model risk assessment Part 1: motivation
- Feiwo technology IPO meeting: annual revenue of 1.13 billion Hunan Cultural Tourism and Yuanli investment are shareholders
猜你喜欢
Can the variable name be in Chinese? Directly fooled people
Doris学习笔记之与其他系统集成
Realize a family security and environmental monitoring system (II)
Alibaba mqtt IOT platform "cloud product circulation" practice - the two esp32 achieve remote interoperability through the IOT platform
Setting of parameter configuration tool for wireless vibrating wire collector
Brush questions - Luogu -p1075 prime factor decomposition
数字孪生 - 认知篇
Doris learning notes integration with other systems
Digital Twins - cognition
Brush questions - Luogu -p1151 sub number integer
随机推荐
Nuc980 set up SSH xshell connection
R语言如何将大型Excel文件转为dta格式详解
Famous handwritten note taking software recruit CTO · coordinate Shenzhen
wangeditor 富文本编辑器
金鱼哥RHCA回忆录：CL210管理存储--管理共享文件系统
依迅总经理孙峰：公司已完成股改，准备IPO
CDA level Ⅰ 2021 new version simulation question 1 (with answers)
机械制造业数字化新“引擎”供应链协同管理系统助力企业精细化管理迈上新台阶
Feiwo technology IPO meeting: annual revenue of 1.13 billion Hunan Cultural Tourism and Yuanli investment are shareholders
Realize a family security and environmental monitoring system (I)
From fish eye to look around to multi task King bombing -- a review of Valeo's classic articles on visual depth estimation (from fisheyedistancenet to omnidet) (Part 2)
Working mode and sleep mode of nlm5 series wireless vibrating wire sensor acquisition instrument
Pytorch uses tensorboard to realize visual summary
CDA level Ⅰ 2021 new version simulation question 2 (with answers)
【目录爆破工具】信息收集阶段：robots.txt、御剑、dirsearch、Dirb、Gobuster
Brush questions - Luogu -p1075 prime factor decomposition
金鱼哥RHCA回忆录：CL210管理存储--对象存储
Leetcode 205. isomorphic string (2022.07.24)
Common problems in the use of wireless vibrating wire acquisition instrument
力扣（LeetCode）205. 同构字符串（2022.07.24）