要使用ARKit进行图像和深度的录制和回放，可以按照以下步骤进行：

1. 创建一个AR会话并启用深度数据：

import ARKit

// 创建AR会话
let arSession = ARSession()

// 启用深度数据
let configuration = ARWorldTrackingConfiguration()
configuration.isDepthDataEnabled = true
arSession.run(configuration)

2. 在AR会话中捕捉图像和深度数据：

extension ViewController: ARSessionDelegate {
    // 实现AR会话委托方法
    func session(_ session: ARSession, didUpdate frame: ARFrame) {
        // 获取图像和深度数据
        guard let image = frame.capturedImage, let depthData = frame.sceneDepth?.depthData else {
            return
        }
        
        // 处理图像和深度数据...
    }
}

3. 创建AVAssetWriter对象来录制图像和深度数据：

import AVFoundation

// 创建AVAssetWriter对象
guard let assetWriter = try? AVAssetWriter(outputURL: outputURL, fileType: .mov) else {
    return
}

// 设置视频输出
let videoOutputSettings = [
    AVVideoCodecKey: AVVideoCodecType.h264,
    AVVideoWidthKey: 1920,
    AVVideoHeightKey: 1080
] as [String : Any]
let videoOutput = AVAssetWriterInput(mediaType: .video, outputSettings: videoOutputSettings)
assetWriter.add(videoOutput)

// 设置深度数据输出
let depthOutputSettings = [
    kCVPixelBufferPixelFormatTypeKey as String: kCVPixelFormatType_DepthFloat32
] as [String : Any]
let depthOutput = AVAssetWriterInput(mediaType: .depthData, outputSettings: depthOutputSettings)
assetWriter.add(depthOutput)

4. 在AR会话委托方法中将图像和深度数据写入AVAssetWriter：

extension ViewController: ARSessionDelegate {
    // 实现AR会话委托方法
    func session(_ session: ARSession, didUpdate frame: ARFrame) {
        // 获取图像和深度数据
        guard let image = frame.capturedImage, let depthData = frame.sceneDepth?.depthData else {
            return
        }
        
        // 将图像和深度数据写入AVAssetWriter
        if assetWriter.status == .unknown {
            assetWriter.startWriting()
            assetWriter.startSession(atSourceTime: frame.timestamp)
        }
        
        if videoOutput.isReadyForMoreMediaData {
            let videoBuffer = CMSampleBufferCreateForImageBuffer(pixelBuffer: image, formatDescription: nil, sampleTiming: CMSampleTimingInfo())
            videoOutput.append(videoBuffer)
        }
        
        if depthOutput.isReadyForMoreMediaData {
            let depthBuffer = CMSampleBufferCreateForImageBuffer(pixelBuffer: depthData, formatDescription: nil, sampleTiming: CMSampleTimingInfo())
            depthOutput.append(depthBuffer)
        }
    }
}

5. 在适当的时机停止录制并保存文件：

// 停止会话
arSession.pause()

// 停止写入数据
videoOutput.markAsFinished()
depthOutput.markAsFinished()
assetWriter.finishWriting {
    // 完成写入数据后的逻辑...
}

请注意，这只是一个简单的示例，供参考。实际使用中可能需要进行更多的错误处理和逻辑控制。