多目标识别

适用场景

可同时检测出给定图片中的各种物体，包括风景、动物、植物、建筑、人脸、表格、文本等位置，并框选出物体。

效果如下图所示：

约束与限制

该能力当前不支持模拟器。

AI能力	约束
多目标识别	- 输入图像具有合适成像的质量（建议720p以上），100px<高度<10000px，100px<宽度<10000px，高宽比例建议5:1以下（高度小于宽度的5倍），接近手机屏幕高宽比例为宜。 - 图片中的物体占比需要大于0.1%。

开发步骤

在使用多目标识别时，将实现多目标识别相关的类添加至工程。

import { image } from '@kit.ImageKit';
import { hilog } from '@kit.PerformanceAnalysisKit';
import { BusinessError } from '@kit.BasicServicesKit';
import { fileIo } from '@kit.CoreFileKit';
import { objectDetection, visionBase } from '@kit.CoreVisionKit';
import { photoAccessHelper } from '@kit.MediaLibraryKit';

简单配置页面的布局，并在Button组件添加点击事件，拉起图库，选择图片。

Button('选择图片')
  .type(ButtonType.Capsule)
  .fontColor(Color.White)
  .alignSelf(ItemAlign.Center)
  .width('80%')
  .margin(10)
  .onClick(() => {
    // 拉起图库，获取图片资源
    void this.selectImage();
  })

通过图库获取图片资源，将图片转换为PixelMap。

private async selectImage() {
  let uri = await this.openPhoto()
  if (uri === undefined) {
    hilog.error(0x0000, 'objectDetectSample', "Failed to define uri.");
  }
  this.loadImage(uri)
}

private async openPhoto(): Promise<string> {
  return new Promise<string>((resolve, reject) => {
    let photoPicker: photoAccessHelper.PhotoViewPicker = new photoAccessHelper.PhotoViewPicker();
    photoPicker.select({
      MIMEType: photoAccessHelper.PhotoViewMIMETypes.IMAGE_TYPE, maxSelectNumber: 1
    }).then(res => {
      resolve(res.photoUris[0])
    }).catch((err: BusinessError) => {
      hilog.error(0x0000, 'objectDetectSample', `Failed to get photo image uri. code: ${err.code}, message: ${err.message}`);
      reject('')
    })
  })
}

private loadImage(name: string) {
  setTimeout(async () => {
    let fileSource = await fileIo.open(name, fileIo.OpenMode.READ_ONLY);
    this.imageSource = image.createImageSource(fileSource.fd);
    this.chooseImage = await this.imageSource.createPixelMap();
  }, 100)
}

实例化Request对象，并传入待检测图片的PixelMap，调用多目标识别的实现多目标识别功能。

// 调用多目标检测接口
let request: visionBase.Request = {
  inputData: { pixelMap: this.chooseImage }
};
let data: objectDetection.ObjectDetectionResponse = await (await objectDetection.ObjectDetector.create()).process(request);

（可选）如果需要将结果展示在界面上，可以使用下列代码。

let objectJson = JSON.stringify(data);
hilog.info(0x0000, 'objectDetectSample', `Succeeded in object detection: ${objectJson}`);
this.dataValues = objectJson;

开发实例

Index.ets

import { image } from '@kit.ImageKit';
import { hilog } from '@kit.PerformanceAnalysisKit';
import { BusinessError } from '@kit.BasicServicesKit';
import { fileIo } from '@kit.CoreFileKit';
import { objectDetection, visionBase } from '@kit.CoreVisionKit';
import { photoAccessHelper } from '@kit.MediaLibraryKit';

@Entry
@Component
struct Index {
  private imageSource: image.ImageSource | undefined = undefined;
  @State chooseImage: PixelMap | undefined = undefined
  @State dataValues: string = ''

  build() {
    Column() {
      Image(this.chooseImage)
        .objectFit(ImageFit.Fill)
        .height('60%')

      Text(this.dataValues)
        .copyOption(CopyOptions.LocalDevice)
        .height('15%')
        .margin(10)
        .width('60%')

      Button('选择图片')
        .type(ButtonType.Capsule)
        .fontColor(Color.White)
        .alignSelf(ItemAlign.Center)
        .width('80%')
        .margin(10)
        .onClick(() => {
          // 拉起图库
          void this.selectImage()
        })

      Button('开始多目标识别')
        .type(ButtonType.Capsule)
        .fontColor(Color.White)
        .alignSelf(ItemAlign.Center)
        .width('80%')
        .margin(10)
        .onClick(() => {
          // 调用封装的异步识别函数
          void this.handleMultiObjectDetection();
        })
    }
    .width('100%')
    .height('100%')
    .justifyContent(FlexAlign.Center)
  }

  // 封装多目标识别的异步逻辑
  private async handleMultiObjectDetection() {
    if(!this.chooseImage) {
      hilog.error(0x0000, 'objectDetectSample', `Failed to choose image.`);
      return;
    }
    let request: visionBase.Request = {
      inputData: { pixelMap: this.chooseImage }
    };
    try {
      let data: objectDetection.ObjectDetectionResponse =
        await (await objectDetection.ObjectDetector.create()).process(request);
      let objectJson = JSON.stringify(data);
      hilog.info(0x0000, 'objectDetectSample', `Succeeded in object detection: ${objectJson}`);
      this.dataValues = objectJson;
    } catch (error) {
      hilog.error(0x0000, 'objectDetectSample', `Failed to get result. Error: ${error}`);
    }
  }

  private async selectImage() {
    try {
      let uri = await this.openPhoto();
      if (uri === undefined) {
        hilog.error(0x0000, 'objectDetectSample', "Failed to define uri.");
        return;
      }
      this.loadImage(uri);
    } catch (err) {
      hilog.error(0x0000, 'objectDetectSample', `Failed to get photo image uri. code: ${err.code}, message: ${err.message}`);
    }
  }

  private async openPhoto(): Promise<string> {
    return new Promise<string>((resolve, reject) => {
      let photoPicker: photoAccessHelper.PhotoViewPicker = new photoAccessHelper.PhotoViewPicker();
      photoPicker.select({
        MIMEType: photoAccessHelper.PhotoViewMIMETypes.IMAGE_TYPE, maxSelectNumber: 1
      }).then(res => {
        resolve(res.photoUris[0]);
      }).catch((err: BusinessError) => {
        hilog.error(0x0000, 'objectDetectSample', `Failed to get photo image uri. code: ${err.code}, message: ${err.message}`);
        reject(err);
      })
    })
  }

  private loadImage(name: string) {
    setTimeout(async () => {
      try {
        let fileSource = await fileIo.open(name, fileIo.OpenMode.READ_ONLY);
        this.imageSource = image.createImageSource(fileSource.fd);
        this.chooseImage = await this.imageSource.createPixelMap();
        await fileIo.close(fileSource);
      } catch (error) {
        hilog.error(0x0000, 'objectDetectSample', `Failed to open file. Error: ${error}`);
      }
    }, 100)
  }
}

适用场景​

约束与限制​

开发步骤​

开发实例​

Index.ets​

适用场景

约束与限制

开发步骤

开发实例

Index.ets