背景

最近打算写一个 macos 翻译软件, 需要用到 ocr 图像识别, 经过翻阅文档和 chatgpt 拉扯了一下午, 最终成功实现.

代码

代码逻辑为, 接受参数: 图片路径, 然后获取图片, 通过 VNImageRequestHandler 对图片进行文字识别

如下代码可以直接放进一个 ocr.swift, 然后执行 swiftc -o ocr ocr.swift , 在执行 ./ocr /Users/yelog/Desktop/3.png 后面为你实际的有文字的图片路经

////  ocr.swift//  Fast Translation////  Created by 杨玉杰 on 2023/12/31.//import SwiftUIimport Visionfunc handleDetectedText(request: VNRequest?, error: Error?) {    if let error = error {        print("ERROR: \(error)")        return    }    guard let results = request?.results, results.count > 0 else {        print("No text found")        return    }    for result in results {        if let observation = result as? VNRecognizedTextObservation {            for text in observation.topCandidates(1) {                let string = text.string                print("识别: \(string)")            }        }    }}func ocrImage(path: String) {    let cgImage = NSImage(byReferencingFile: path)?.ciImage()?.cgImage    let requestHandler = VNImageRequestHandler(cgImage: cgImage!)    let request = VNRecognizeTextRequest(completionHandler: handleDetectedText)    // 设置文本识别的语言为英文    request.recognitionLanguages = ["en"]    request.recognitionLevel = .accurate    do {        try requestHandler.perform([request])    } catch {        print("Unable to perform the requests: \(error).")    }}extension NSImage {    func ciImage() -> CIImage? {        guard let data = self.tiffRepresentation,              let bitmap = NSBitmapImageRep(data: data) else {            return nil        }        let ci = CIImage(bitmapImageRep: bitmap)        return ci    }}// 执行函数，从命令行参数中获取图片的地址ocrImage(path: CommandLine.arguments[1])

然后准备待识别的有文字的图片

# 编译 swift 文件swiftc -o ocr ocr.swift# 执行并且传递图片路径参数./ocr /Users/yelog/Desktop/3.png

最后

最近打算着手写一些 macos 的小工具, 如果对 swift 或者 macos 感兴趣的可以关注或评论.