如何用Swift同时生成多个RealityKit模型？

单模型生成基础代码 #

原本生成单个模型的代码如下，通过指定图片路径、输出路径和细节等级等参数调用HelloPhotogrammetry.main()：

HelloPhotogrammetry.main([path_to_folder_of images, path_to_output/model.usdz, "-d", "medium", "-o", "unordered", "-f", "high" ])

批量生成需求 #

需要同时生成3个仅输出名称和细节等级不同的模型，对应参数如下：

• 输出文件：modelname_medium.usdz，细节等级参数：-d medium
• 输出文件：modelname_full.usdz，细节等级参数：-d full
• 输出文件：modelname_raw.usdz，细节等级参数：-d raw

你的循环代码可行性分析 #

你写的循环代码逻辑上是可行的，但存在几个需要修正的细节：

1. 字符串换行拼接可能导致Swift语法错误（当前代码里的路径被拆成多行，未正确拼接）
2. 你定义的levelOfDetail数组包含reduced，但需求里是raw，需要保持参数一致
3. 硬编码完整路径不利于后续维护和修改

修正后的可运行代码示例：

// 修正细节等级数组，匹配需求
let levelOfDetail = ["medium", "full", "raw"]
// 提取固定路径为常量，简化代码
let imageFolderPath = "/Users/paulmmxx/Desktop/Object_Capture_API/dataset/jpeg/all"
let outputBasePath = "/Users/paulmmxx/Desktop/Object_Capture_API/output/preiser_"

for detail in levelOfDetail {
    let outputPath = outputBasePath + detail + ".usdz"
    HelloPhotogrammetry.main([
        imageFolderPath,
        outputPath,
        "-d", detail,
        "-o", "unordered",
        "-f", "high"
    ])
}

更优实现方案 #

1. 用元组关联细节等级与输出标识 #

如果后续需要调整输出文件名后缀或参数，用元组存储对应关系更清晰，便于维护：

// 用元组存储细节等级和对应的输出文件名后缀
let detailConfigs = [
    ("medium", "medium"),
    ("full", "full"),
    ("raw", "raw")
]

let imageFolderPath = "/Users/paulmmxx/Desktop/Object_Capture_API/dataset/jpeg/all"
let outputDirPath = "/Users/paulmmxx/Desktop/Object_Capture_API/output/"

for (detailLevel, suffix) in detailConfigs {
    let outputPath = outputDirPath + "preiser_" + suffix + ".usdz"
    HelloPhotogrammetry.main([
        imageFolderPath,
        outputPath,
        "-d", detailLevel,
        "-o", "unordered",
        "-f", "high"
    ])
}

这种方式可以灵活调整文件名后缀，比如想把raw对应的文件名改成preiser_highres.usdz，只需要修改元组即可。

2. 增加基础路径的有效性检查 #

可以提前检查图片输入路径和输出目录是否存在，避免运行时出错：

import Foundation

let imageFolderPath = "/Users/paulmmxx/Desktop/Object_Capture_API/dataset/jpeg/all"
let outputDirPath = "/Users/paulmmxx/Desktop/Object_Capture_API/output/"

// 检查输入路径是否存在
guard FileManager.default.fileExists(atPath: imageFolderPath) else {
    print("错误：图片文件夹路径不存在")
    exit(1)
}

// 检查输出目录是否存在，不存在则创建
if !FileManager.default.fileExists(atPath: outputDirPath) {
    do {
        try FileManager.default.createDirectory(atPath: outputDirPath, withIntermediateDirectories: true)
    } catch {
        print("创建输出目录失败：\(error.localizedDescription)")
        exit(1)
    }
}

let detailConfigs = [("medium", "medium"), ("full", "full"), ("raw", "raw")]

for (detailLevel, suffix) in detailConfigs {
    let outputPath = outputDirPath + "preiser_" + suffix + ".usdz"
    HelloPhotogrammetry.main([
        imageFolderPath,
        outputPath,
        "-d", detailLevel,
        "-o", "unordered",
        "-f", "high"
    ])
}

3. 异步执行（可选） #

如果HelloPhotogrammetry.main()是同步阻塞调用，批量生成时会耗时较长。若API支持多线程调用，可以用DispatchQueue并行处理，提升效率：

import Foundation

let imageFolderPath = "/Users/paulmmxx/Desktop/Object_Capture_API/dataset/jpeg/all"
let outputDirPath = "/Users/paulmmxx/Desktop/Object_Capture_API/output/"

// 提前完成路径检查（代码同方案2）
guard FileManager.default.fileExists(atPath: imageFolderPath) else {
    print("错误：图片文件夹路径不存在")
    exit(1)
}
if !FileManager.default.fileExists(atPath: outputDirPath) {
    do {
        try FileManager.default.createDirectory(atPath: outputDirPath, withIntermediateDirectories: true)
    } catch {
        print("创建输出目录失败：\(error.localizedDescription)")
        exit(1)
    }
}

let detailConfigs = [("medium", "medium"), ("full", "full"), ("raw", "raw")]
let queue = DispatchQueue(label: "com.example.objectcapture", attributes: .concurrent)
let group = DispatchGroup()

for (detailLevel, suffix) in detailConfigs {
    group.enter()
    queue.async {
        let outputPath = outputDirPath + "preiser_" + suffix + ".usdz"
        HelloPhotogrammetry.main([
            imageFolderPath,
            outputPath,
            "-d", detailLevel,
            "-o", "unordered",
            "-f", "high"
        ])
        group.leave()
    }
}

// 等待所有任务完成
group.wait()
print("所有模型生成完成")

内容的提问来源于stack exchange，提问作者Paul St George