Skip to content

【场景】图片懒加载 ✨

要点速览

  • IntersectionObserver:检测元素与视口/祖先容器相交,驱动懒加载与曝光
  • 关键参数:root/rootMargin/threshold;事件项含 isIntersecting/target/intersectionRatio
  • 懒加载原理:初始占位图 + 真实地址存于 data-src,进入可视区域后替换为 src
  • 方案选择:现代优先 IntersectionObserver;旧浏览器用滚动事件并节流/缓冲
  • 提前加载:用 rootMargin 扩大底部判定区以预取(如 100px
  • 性能与可访问性:加载后 unobserve;错误兜底;避免布局抖动与过度动画

IntersectionObserver

IntersectionObserver 是一个现代浏览器 API,用于检测一个元素(或其子元素)相对于视口或某个祖先元素的可见性变化

基本用法

js
const ob = new IntersectionObserver(callback, options);
  1. callback(entries, observer): 当被观察元素的可见性变化时调用的回调函数callback 一开始会触发一次,确认当前的可视状态(无论当前是可见还是不可见),之后在每次可视状态发生改变时会触发。回调函数里面有两个参数:

    • entries: 一个数组,包含所有被观察元素的 IntersectionObserverEntry 对象,每个对象包含以下属性:
      • boundingClientRect: 被观察元素的矩形区域信息。
      • intersectionRatio: 被观察元素的可见部分与整个元素的比例。
      • intersectionRect: 可见部分的矩形区域信息。
      • isIntersecting: 布尔值,表示元素是否与根元素相交
      • rootBounds: 根元素的矩形区域信息。
      • target: 被观察的目标元素
      • time: 触发回调的时间戳。
    • observer: IntersectionObserver 实例本身。
  2. options: 配置对象,用于定制观察行为

    • root指定用作视口的元素,默认值为 null,表示使用浏览器视口作为根元素。

    • rootMargin: 类似于 CSS 的 margin 属性,定义根元素的外边距,用于扩展或缩小根元素的判定区域。可以用像素或百分比表示,例如 '10px' 或 '10%'。

    • threshold: 是一个 0 ~ 1 之间的值,表示一个触发的阈值,如果是 0,只要目标元素一碰到 root 元素,就会触发,如果是 1,表示目标元素完全进入 root 元素范围,才会触发。设置观察元素进入到根元素的百分比。

有了 observer 实例对象后,要观察哪个元素,直接通过 observe 方法来进行观察即可,取消观察通过 unobserve 方法:

js
// 开始观察
ob.observe(elementA);
ob.observe(elementB);

// 停止观察
ob.unobserve(element);

示例 1

html
<!DOCTYPE html>
<html lang="en">
  <head>
    <meta charset="UTF-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <title>Document</title>
    <style>
      body {
        height: 2000px; /*便于出现滚动条*/
        margin: 0;
        padding: 0;
      }

      .target {
        width: 200px;
        height: 200px;
        background-color: #e81b3a;
        margin: 0 auto;
        margin-top: 1500px;
      }
    </style>
  </head>
  <body>
    <div class="target"></div>
    <script>
      // 获取被观察的元素
      const target = document.querySelector(".target");

      // 当被观察的元素可见性发生变化的时候会调用这个函数
      function callback(entries) {
        // 检查目标元素是否进入视口
        entries.forEach((element) => {
          if (element.isIntersecting) {
            console.log("进入视口");
          } else {
            console.log("离开视口");
          }
        });
      }

      // 创建观察器
      const ob = new IntersectionObserver(callback, {
        root: null, // 将浏览器视口作为根元素
        rootMargin: "0px", // 根元素的外边距
        threshold: 0, // 目标元素进入到根元素的百分比
      });

      // 观察目标元素
      ob.observe(target);
    </script>
  </body>
</html>

示例 2

html
<!DOCTYPE html>
<html lang="en">
  <head>
    <meta charset="UTF-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <title>Document</title>
    <style>
      .container {
        width: 100%;
        height: 500px;
        border: 1px solid black;
        overflow: scroll;
      }

      .target {
        width: 200px;
        height: 200px;
        background-color: lightcoral;
        margin: 0 auto;
        margin-top: 1000px;
      }
    </style>
  </head>
  <body>
    <div class="container">
      <div class="target"></div>
    </div>
    <script>
      // 获取被观察的元素
      const target = document.querySelector(".target");

      // 当被观察的元素可见性发生变化的时候会调用这个函数
      function callback(entries) {
        // 检查目标元素是否进入视口
        entries.forEach((element) => {
          if (element.isIntersecting) {
            console.log("进入视口");
          } else {
            console.log("离开视口");
          }
        });
      }

      // 创建观察器
      const ob = new IntersectionObserver(callback, {
        root: document.querySelector(".container"), // 将 container 作为根元素
        rootMargin: "-50px", // 根元素的外边距(负数表示向内收缩)
        threshold: 0, // 目标元素进入到根元素的百分比
      });

      // 观察目标元素
      ob.observe(target);
    </script>
  </body>
</html>

懒加载

懒加载含义:当出现的时候再加载。

懒加载核心原理

懒加载核心原理:img 元素在 src 属性有值时,才会去请求对应的图片地址,那么我们可以先给图片一张默认的占位图:

html
<img src="占位图.png" />

再设置一个自定义属性 data-src,对应的值为真实的图片地址:

html
<img src="占位图.png" data-src="图片真实地址" />

之后判断当然这个 img 元素有没有进入可视区域,如果进入了,就把 data-src 的值赋给 src,让真实的图片显示出来。这就是图片懒加载的基本原理。

不过这里对于判断 img 元素有没有进入可视区域,有着新旧两套方案。

旧方案

早期的方案是监听页面的滚动:

js
window.addEventListener("scroll", () => {});

当 img 标签的顶部到可视区域顶部的距离,小于可视区域高度的时候,我们就认为图片进入了可视区域,画张图表示:

image-20240724154242876

示例代码:

js
window.addEventListener("scroll", () => {
  const img = document.querySelectorAll("img");
  img.forEach((img) => {
    const rect = img.getBoundingClientRect();
    if (rect.top < document.body.clientHeight) {
      // 当前这张图片进入到可视区域
      // 做 src 的替换
      img.src = img.dataset.src;
    }
  });
});

缺点:

  • scroll 事件频繁触发
  • 需要手动计算目标元素是否进入可视区域

新方案

使用 IntersectionObserver 来实现。

js
let observer = new IntersectionObserver(
  (entries, observer) => {
    for (const entry of entries) {
      if (entry.isIntersecting) {
        // 进入此分支,说明当前的图片和根元素产生了交叉
        const img = entry.target; // 拿到观察的目标元素
        img.src = img.dataset.src; // 属性替换
        observer.unobserve(img); // 取消观察
      }
    }
  },
  {
    root: null, // 将浏览器视口作为根元素
    rootMargin: "0px 0px 0px 0px",
    threshold: 0.5,
  }
);
// 先拿到所有的图片元素
const imgs = document.querySelectorAll("img");
imgs.forEach((img) => {
  //观察所有的图片元素
  observer.observe(img);
});

与旧方案相比性能更高,只有在目标元素和根元素发生交叉的时候才会触发回调。

结合两种方案

js
// 优化后的懒加载方案
document.addEventListener("DOMContentLoaded", () => {
  const lazyImages = Array.from(document.querySelectorAll("img[data-src]"));
  if (!lazyImages.length) return;

  // 方案1:优先使用 IntersectionObserver(现代浏览器)
  if ("IntersectionObserver" in window) {
    const observer = new IntersectionObserver(
      (entries, observer) => {
        entries.forEach((entry) => {
          if (entry.isIntersecting) {
            const img = entry.target;
            img.src = img.dataset.src;
            img.onload = () => img.removeAttribute("data-src");
            observer.unobserve(img);
          }
        });
      },
      {
        rootMargin: "0px 0px 100px 0px", // 提前100px加载
      }
    );

    lazyImages.forEach((img) => observer.observe(img));
  }
  // 方案2:兼容旧浏览器的滚动方案
  else {
    let active = false; // 标记是否函数正在运行
    const lazyLoad = () => {
      if (active) return;
      active = true;

      setTimeout(() => {
        lazyImages.forEach((img) => {
          const rect = img.getBoundingClientRect();
          if (rect.top <= window.innerHeight + 100) {
            // +100px缓冲
            img.src = img.dataset.src;
            img.onload = () => img.removeAttribute("data-src");
            let index = lazyImages.indexOf(img);
            if (index !== -1) {
              lazyImages.splice(index, 1); // 去除加载完成的图片
            }
          }
        });
        active = false;
      }, 200);
    };

    // 初始化加载可视区图片
    lazyLoad();

    // 滚动事件监听(带节流)
    window.addEventListener("scroll", lazyLoad);
    window.addEventListener("resize", lazyLoad);
  }
});

在线演示

Vue 相关库

在 Vue 中 vue3-observe-visibility 库对 IntersectionObserver API 进行了进一步封装实现的

安装:

bash
npm install --save vue3-observe-visibility

注册

js
import { createApp } from "vue";
import App from "./App.vue";
// 引入该第三方库
import { ObserveVisibility } from "vue3-observe-visibility";

const app = createApp(App);

// 将其注册成为一个全局的指令
app.directive("observe-visibility", ObserveVisibility);

app.mount("#app");

使用示例

vue
<template>
  <div>
    <h1>示例</h1>
    <!-- 该元素进入或者离开视口时,会触发回调函数 -->
    <div
      v-observe-visibility="{
        callback: visibilityChange,
        intersection: {
          root: null,
          rootMargin: '0px',
          threshold: 0,
        }, // 配置对象
      }"
      class="target"
    ></div>
  </div>
</template>

<script setup>
function visibilityChange(isVisible) {
  console.log(isVisible ? "进入视口" : "离开视口");
}
</script>

<style scoped>
.target {
  height: 200px;
  width: 200px;
  background-color: pink;
  margin: 0 auto;
  margin-top: 1000px;
}
</style>

实战演练

js
import { createApp } from "vue";
import App from "./App.vue";
import { ObserveVisibility } from "vue3-observe-visibility";

const app = createApp(App);
// 注册指令
app.directive("observe-visibility", ObserveVisibility);

app.mount("#app");
vue
<template>
  <div>
    <h1>图片懒加载示例</h1>
    <div class="image-grid">
      <!-- 一定要配置 once 配置项 -->
      <!-- 否则会在可视状态发生变化时反复加载 -->
      <img
        v-observe-visibility="{
          callback: visibilityChanged,
          once: true, // 相当于加载一次之后调用 unobserve
          intersection: {
            root: null,
            rootMargin: '0px',
            threshold: 0.1,
          },
        }"
        v-for="(url, index) in imageUrls"
        :key="index"
        :data-src="url"
        :alt="'id ' + (index + 1)"
        :src="loadingImage"
        @error="handleError"
      />
    </div>
  </div>
</template>

<script setup>
import { ref } from "vue";

// 生成一些图片URL
const imageUrls = ref([]);
// 往 imageUrls 中添加 50 个图片 URL
for (let i = 1; i <= 50; i++) {
  imageUrls.value.push(`https://picsum.photos/id/${i}/600/400`);
}

// 加载图片的 url
const loadingImage =
  "https://dummyimage.com/600x400/cccccc/000000&text=Loading";
// 错误图片的 url
const errorImage = "https://dummyimage.com/600x400/ff0000/ffffff&text=Error";

function visibilityChanged(visibility, entry) {
  const img = entry.target;
  if (visibility) {
    img.src = img.dataset.src;
  }
}

// 图片加载失败时的处理函数
function handleError(event) {
  const img = event.target;
  img.src = errorImage;
}
</script>

<style scoped>
.image-grid {
  display: flex;
  flex-wrap: wrap;
  width: 800px;
  margin: 0 auto;
}

.image-grid img {
  display: block;
  margin: 10px;
  width: 200px;
  height: 150px;
  object-fit: cover;
}
</style>