Q4: 什么是射线投射?在 VR 看房项目中如何应用?
口头回答
面试官您好,射线投射(Raycasting)是3D图形学中的一项重要技术,它的基本原理是从一个点(通常是相机位置)沿着特定方向发射一条虚拟射线,然后检测这条射线与场景中哪些3D物体相交。
射线投射的核心概念:
- 射线定义:由起点和方向向量定义的无限延伸的直线
- 相交检测:计算射线与3D物体(如网格、精灵等)的交点
- 距离排序:按照交点到射线起点的距离进行排序
- 交互响应:根据相交结果执行相应的交互逻辑
在我们的VR看房项目中,射线投射主要应用于两个核心场景:
导航点击交互:当用户点击房间导航图标时,通过射线投射检测点击的是哪个导航精灵,然后执行相应的房间切换动画。
信息点悬停检测:当用户鼠标悬停在信息点上时,通过射线投射实时检测鼠标位置对应的信息点,并显示相应的提示信息。
这种技术的优势在于能够在3D空间中实现精确的交互检测,无需复杂的碰撞检测算法,性能高效且实现简单。
项目中的具体实现细节
1. 射线投射基础设置
// 在 App.vue 中创建射线投射器和指针向量
let pointer = new THREE.Vector2(); // 存储归一化的鼠标坐标
let raycaster = new THREE.Raycaster(); // 射线投射器对象
2. 鼠标坐标归一化处理
// 将屏幕坐标转换为Three.js的标准化设备坐标
function normalizeMouseCoordinates(event: MouseEvent) {
// X坐标:从 [0, window.innerWidth] 映射到 [-1, 1]
pointer.x = (event.clientX / window.innerWidth) * 2 - 1;
// Y坐标:从 [0, window.innerHeight] 映射到 [1, -1]
// 注意Y轴需要翻转,因为屏幕坐标Y轴向下,而Three.js坐标Y轴向上
pointer.y = -(event.clientY / window.innerHeight) * 2 + 1;
}
3. 导航点击交互实现(RoomSprite.ts)
// 在 RoomSprite 类中实现点击检测
export class RoomSprite {
sprite: THREE.Sprite;
callbacks: any[]; // 存储点击回调函数
constructor(
text: string,
position: THREE.Vector3,
scene: THREE.Scene,
camera: THREE.Camera
) {
this.callbacks = [];
// 创建导航精灵...
// 监听全局点击事件
window.addEventListener("click", (event) => {
// 1. 归一化鼠标坐标
pointer.x = (event.clientX / window.innerWidth) * 2 - 1;
pointer.y = -(event.clientY / window.innerHeight) * 2 + 1;
// 2. 设置射线起点和方向
raycaster.setFromCamera(pointer, camera);
// 3. 检测射线与当前精灵的相交
const intersects = raycaster.intersectObject(sprite);
// 4. 如果有相交点,执行回调函数
if (intersects.length > 0) {
this.callbacks.forEach((callback) => callback());
}
});
}
// 添加点击回调函数
onClick(callback: () => void) {
this.callbacks.push(callback);
}
}
4. 信息点悬停检测实现(App.vue)
// 存储所有信息点精灵的数组
const spriteList: THREE.Sprite[] = [];
// 信息点悬停显示函数
function tooltipShow(e: MouseEvent, spriteList: THREE.Sprite[]) {
e.preventDefault();
const pointer = new THREE.Vector2();
const raycaster = new THREE.Raycaster();
// 1. 归一化鼠标坐标
pointer.x = (e.clientX / window.innerWidth) * 2 - 1;
pointer.y = -(e.clientY / window.innerHeight) * 2 + 1;
// 2. 从相机位置发射射线
raycaster.setFromCamera(pointer, camera);
// 3. 检测射线与所有信息点的相交
const intersects = raycaster.intersectObjects(spriteList);
// 4. 处理相交结果
if (intersects.length > 0) {
// 检查是否为信息点类型
if (intersects[0].object.userData.type === "information") {
// 计算提示框在屏幕上的位置
const element = e.target as HTMLElement;
const elementWidth = element.clientWidth / 2;
const elementHeight = element.clientHeight / 2;
// 将3D世界坐标转换为屏幕坐标
const worldVector = new THREE.Vector3(
intersects[0].object.position.x,
intersects[0].object.position.y,
intersects[0].object.position.z
);
const position = worldVector.project(camera);
// 计算提示框的屏幕位置
const left = Math.round(
(position.x + 1) * elementWidth - tooltipBox.value!.clientWidth / 2
);
const top = Math.round(
-(position.y - 1) * elementHeight - tooltipBox.value!.clientHeight / 2
);
// 更新提示框位置和内容
tooltipPosition.value = {
left: `${left}px`,
top: `${top}px`,
};
tooltipContent.value = intersects[0].object.userData;
} else {
// 如果不是信息点,隐藏提示框
handleTooltipHide(e);
}
}
}
5. 信息点创建和注册
// 创建信息点并添加到检测列表
const infoSprite1 = new InfoSprite(
"./images/dot.png",
new THREE.Vector3(1.4, -0.2, -3),
scene,
{
type: "information",
name: "画",
description: "这是一幅画",
}
);
const infoSprite2 = new InfoSprite(
"./images/dot.png",
new THREE.Vector3(-2.4, 0.6, -3),
scene,
{
type: "information",
name: "木雕艺术品",
description: "这是一件木雕艺术品",
}
);
const infoSprite3 = new InfoSprite(
"./images/dot.png",
new THREE.Vector3(1.5, 0.4, 0.9),
scene,
{
type: "information",
name: "艺术画",
description: "这是一幅艺术画",
},
0.1
);
// 将所有信息点添加到检测列表
spriteList.push(infoSprite1.sprite);
spriteList.push(infoSprite2.sprite);
spriteList.push(infoSprite3.sprite);
6. 事件监听器注册
// 注册鼠标移动事件监听器
renderer.domElement.addEventListener("mousemove", (e) => {
tooltipShow(e, spriteList);
});
// 注册鼠标离开事件监听器
tooltipBox.value?.addEventListener("mouseleave", handleTooltipHide);
射线投射技术深度分析
1. 坐标系转换原理
// 屏幕坐标到标准化设备坐标的转换
/*
* 屏幕坐标系:
* - 原点在左上角
* - X轴向右为正,范围 [0, window.innerWidth]
* - Y轴向下为正,范围 [0, window.innerHeight]
*
* 标准化设备坐标系(NDC):
* - 原点在屏幕中心
* - X轴向右为正,范围 [-1, 1]
* - Y轴向上为正,范围 [-1, 1]
*/
// X坐标转换:[0, width] -> [-1, 1]
pointer.x = (event.clientX / window.innerWidth) * 2 - 1;
// Y坐标转换:[0, height] -> [1, -1](注意翻转)
pointer.y = -(event.clientY / window.innerHeight) * 2 + 1;
2. 射线生成机制
// raycaster.setFromCamera() 的内部原理
/*
* 1. 根据相机类型(透视/正交)计算射线起点
* 2. 根据鼠标位置和相机参数计算射线方向
* 3. 对于透视相机:
* - 起点:相机位置
* - 方向:从相机位置指向屏幕上鼠标对应的3D点
*/
raycaster.setFromCamera(pointer, camera);
3. 相交检测算法
// intersectObjects() 的工作原理
/*
* 1. 遍历所有目标对象
* 2. 对每个对象执行射线-几何体相交测试
* 3. 计算相交点的距离
* 4. 按距离排序返回结果
*
* 返回的 intersects 数组包含:
* - object: 相交的3D对象
* - distance: 相交点到射线起点的距离
* - point: 相交点的3D坐标
* - face: 相交的面(对于网格对象)
* - uv: 相交点的UV坐标
*/
const intersects = raycaster.intersectObjects(spriteList);
4. 3D到2D坐标投影
// 将3D世界坐标转换为屏幕坐标
const worldVector = new THREE.Vector3(
intersects[0].object.position.x,
intersects[0].object.position.y,
intersects[0].object.position.z
);
// project() 方法的作用:
// 1. 应用相机的视图矩阵
// 2. 应用相机的投影矩阵
// 3. 执行透视除法
// 4. 返回标准化设备坐标 [-1, 1]
const position = worldVector.project(camera);
// 转换为屏幕像素坐标
const left = Math.round(
(position.x + 1) * elementWidth - tooltipBox.value!.clientWidth / 2
);
const top = Math.round(
-(position.y - 1) * elementHeight - tooltipBox.value!.clientHeight / 2
);
性能优化策略
1. 对象筛选优化
// 只对需要交互的对象进行射线检测
// 避免检测所有场景对象,提高性能
const intersects = raycaster.intersectObjects(spriteList); // 只检测精灵
// 而不是:raycaster.intersectObjects(scene.children, true); // 检测所有对象
2. 事件节流
// 在实际项目中可以添加节流,避免过于频繁的检测
let lastCheckTime = 0;
const CHECK_INTERVAL = 16; // 约60FPS
function throttledTooltipShow(e: MouseEvent, spriteList: THREE.Sprite[]) {
const now = Date.now();
if (now - lastCheckTime > CHECK_INTERVAL) {
tooltipShow(e, spriteList);
lastCheckTime = now;
}
}
3. 距离优化
// 可以设置最大检测距离,避免检测过远的对象
raycaster.far = 100; // 只检测100单位内的对象
技术亮点和最佳实践
1. 类型安全的用户数据
// 在 InfoSprite 中使用 userData 存储类型信息
infoSprite.userData = {
type: "information", // 用于区分对象类型
name: "画",
description: "这是一幅画",
};
// 在检测时进行类型判断
if (intersects[0].object.userData.type === "information") {
// 处理信息点逻辑
}
2. 回调函数模式
// RoomSprite 使用回调函数模式,提高代码复用性
export class RoomSprite {
callbacks: any[];
onClick(callback: () => void) {
this.callbacks.push(callback);
}
}
// 使用时可以灵活添加不同的点击行为
balconySprite.onClick(() => {
gsap.to(camera.position, { x: 0, y: 0, z: -10, duration: 1 });
});
3. 事件委托优化
// 使用全局事件监听器而不是为每个对象单独添加
// 这样可以减少事件监听器的数量,提高性能
window.addEventListener("click", handleGlobalClick);
renderer.domElement.addEventListener("mousemove", handleGlobalMouseMove);
应用场景扩展
1. 可能的扩展应用
- 物体高亮:鼠标悬停时高亮显示可交互对象
- 拖拽操作:实现3D对象的拖拽移动
- 碰撞检测:检测移动对象与静态对象的碰撞
- 选择框:实现框选多个3D对象
- 测距工具:点击两点测量3D空间距离
2. 移动端适配
// 触摸事件的射线投射
function handleTouch(event: TouchEvent) {
const touch = event.touches[0];
pointer.x = (touch.clientX / window.innerWidth) * 2 - 1;
pointer.y = -(touch.clientY / window.innerHeight) * 2 + 1;
raycaster.setFromCamera(pointer, camera);
// 执行相交检测...
}
技术总结
射线投射技术在我们的VR看房项目中发挥了关键作用:
- 精确交互:实现了像素级精确的3D空间交互检测
- 性能高效:相比复杂的碰撞检测算法,射线投射计算简单高效
- 易于实现:Three.js提供了完善的射线投射API,实现简单
- 扩展性强:可以轻松扩展到更多交互场景
- 用户体验:提供了直观自然的3D空间交互体验
这种技术的成功应用,使我们的VR看房应用具备了专业级的交互体验,是现代3D Web应用的核心技术之一。