HarmonyOS鸿蒙Next中使用OCR如何识别从沙箱读取的图片？

开发者你好，参考以下方案：

【解决方案】

1. 将本地图片存入沙箱，示例如下：

   let context: Context | undefined= this.getUIContext().getHostContext()
     if (!context) {
     return
   }
   let filePath = context.cacheDir + '/image1.jpg'
   // 读取本地图片
   let arrayBuff =
     context.resourceManager.getRawFileContentSync('Screenshot_2025-04-14T191059.png').buffer 
   let fileSource = await fileIo.open(filePath, fs.OpenMode.READ_WRITE | fs.OpenMode.CREATE);
   // 写入沙箱
   fs.write(fileSource.fd, arrayBuff)
   

2. 将存入沙箱的图片，转换为PixelMap格式，示例如下：

     imageSource = image.createImageSource(filePath);
     // 将图片格式转为PixelMap
     chooseImage = await imageSource.createPixelMap();
   

3. 通过textRecognition.init初始化OCR服务。

4. 识别图片，识别成功后获取到图片上的文字，示例如下：

   textRecognition.recognizeText(visionInfo,
     (error: BusinessError, data: textRecognition.TextRecognitionResult) => {
         if (error.code !== 0) {
               hilog.error(0x0000, 'OCRDemo', `Failed to recognize text. Code: ${error.code}, message: ${error.message}`);
                 return;
          }
      // 识别成功，获取对应的结果
               let recognitionString = JSON.stringify(data.value);
               this.imageText = recognitionString
               hilog.info(0x0000, 'OCRDemo', `Succeeded in recognizing text: ${recognitionString}`);
         if(chooseImage && imageSource) {
               chooseImage.release();
               imageSource.release();
         }
   });
   

完整代码：

import { textRecognition } from '@kit.CoreVisionKit'
import { image } from '@kit.ImageKit';
import { hilog } from '@kit.PerformanceAnalysisKit';
import { BusinessError } from '@kit.BasicServicesKit';
import { fileIo } from '@kit.CoreFileKit';
import { fileIo as fs, fileUri } from '@kit.CoreFileKit';
import { common, Context } from '@kit.AbilityKit';

let imageSource: image.ImageSource | undefined = undefined;
let chooseImage: PixelMap | undefined = undefined;
@Entry
@Component
struct Page {
  @State imageText:string = ''
  
  build() {
    Column(){
      Button('点击一下')
        .width('60%')
        .height(40)
        .backgroundColor(Color.Orange)
        .align(Alignment.Center)
        .onClick(()=>{
          setTimeout(async () => {
            let context: Context | undefined= this.getUIContext().getHostContext()
            if (!context) {
              return
            }
            let filePath = context.cacheDir + '/image1.jpg'
            // 读取本地图片
            let arrayBuff =
              context.resourceManager.getRawFileContentSync('Screenshot_2025-04-14T191059.png').buffer 
            let fileSource = await fileIo.open(filePath, fs.OpenMode.READ_WRITE | fs.OpenMode.CREATE);
            // 写入沙箱
            fs.write(fileSource.fd, arrayBuff)
            imageSource = image.createImageSource(filePath);
            // 将图片格式转为PixelMap
            chooseImage = await imageSource.createPixelMap();
            if (!chooseImage) {
              return;
            }
            // 初始化OCR服务
            const initResult = await textRecognition.init();
            if (initResult) {
              hilog.info(0x0000, 'textRecognitionSample', 'OCR service initialized successfully');
              let visionInfo: textRecognition.VisionInfo = {
                pixelMap: chooseImage
              };
              textRecognition.recognizeText(visionInfo,
                (error: BusinessError, data: textRecognition.TextRecognitionResult) => {
                if (error.code !== 0) {
                  hilog.error(0x0000, 'OCRDemo', `Failed to recognize text. Code: ${error.code}, message: ${error.message}`);
                  return;
                }
                // 识别成功，获取对应的结果
                let recognitionString = JSON.stringify(data.value);
                this.imageText = recognitionString
                hilog.info(0x0000, 'OCRDemo', `Succeeded in recognizing text: ${recognitionString}`);

                if(chooseImage && imageSource) {
                  chooseImage.release();
                  imageSource.release();
                }
              });
              textRecognition.getSupportedLanguages().then((data: Array<string>) => {
                let languageString = data.join(', ');
                hilog.info(0x0000, 'OCRDemo', `Succeeded in obtaining the language: ${languageString}`);
              }, (err: BusinessError) => {
                hilog.error(0x0000, 'OCRDemo', `Failed to obtain the language. Code: ${err.code}, message: ${err.message}`);
              });
              // 使用完毕后，释放OCR服务
              await textRecognition.release();
              hilog.info(0x0000, 'textRecognitionSample', 'OCR service released successfully');
            } else {
              hilog.error(0x0000, 'textRecognitionSample', 'Failed to initialize OCR service');
            }
            if (fileSource) {
              try {
                await fileIo.close(fileSource);
              } catch (err) {
                hilog.error(0x0000, 'OCRDemo', `Failed to close fileSource. Code: ${err.code}，message: ${err.message}`);
              }
            }
          }, 100)
        })
        .margin({bottom:60})
      Text(this.imageText)
        .fontSize(20)
        .fontWeight(FontWeight.Bold)
    }

    .width('100%')
    .height('100%')
    .justifyContent(FlexAlign.Center)
  }
}

【背景知识】
待识别的视觉信息，目前仅支持颜色数据格式为RGBA_8888的PixelMap类型的视觉信息。
被识别的图片约束与限制为：

1. 支持的图片格式：JPEG、JPG、PNG。
2. 支持的语言：简体中文、英文、日文、韩文、繁体中文。
3. 文本长度：不超过10000字符。
4. 支持文档印刷体识别，在识别手写字体方面能力有所欠缺。
5. 输入图像具有合适成像的质量（建议720p以上），100px<高度<15210px，100px<宽度<10000px，高宽比例建议10:1以下（高度小于宽度的10倍），接近手机屏幕高宽比例为宜。
6. 拍摄角度与文本所在平面垂直方向的夹角应小于30度。