{ "article_title": "ADG t_img 图片翻译接口全解析：从入门到工程化实战", "article_topic": "图片翻译API接口技术指南", "tags": ["图片翻译", "OCR", "API接口", "多语言", "工程化"], "summary": "t_img 是万邦开放平台推出的图片翻译专用接口，集 OCR 文字识别与机器翻译于一体，支持 20+ 语种互译，兼容 JPG/PNG 等主流格式，可保持原图排版布局，广泛应用于跨境电商、文档翻译、截图实时翻译等场景。本文从接口原理、参数详解、返回结构到代码实战，系统梳理 t_img 接口的完整接入方案。", "sections": [ { "section_title": "一、接口概述", "content": "t_img 是图片翻译专用 API 接口，实现了「OCR 文字识别 + 机器翻译 + 版式回填」一体化处理。开发者只需传入图片 URL 或 Base64 编码，指定目标语言，即可一次性获取识别文本、翻译结果以及文字坐标信息，无需手动拼接 OCR 和翻译两个接口。", "key_features": [ "OCR + 翻译一体化，单次调用完成全流程", "支持 20+ 语种互译（中、英、日、韩、法、德、俄、西等）", "兼容 JPG、PNG、BMP、WEBP 等主流图片格式", "返回文字区域坐标（rect/points），支持版式还原", "支持图片 URL 和 Base64 两种传入方式", "内置缓存机制，重复请求响应更快" ] }, { "section_title": "二、接口信息与请求参数", "content": "接口采用 HTTP GET/POST 方式调用，公共参数通过 URL 拼接，业务参数支持 GET 查询串或 POST JSON Body。", "request_info": { "api_name": "translate.t_img", "endpoint": "https://api-gw.onebound.cn/translate/t_img", "method": "GET / POST", "content_type": "application/x-www-form-urlencoded / application/json" }, "public_params": [ {"name": "key", "type": "String", "required": true, "desc": "调用密钥 Key，GET 方式拼接在 URL 中"}, {"name": "secret", "type": "String", "required": true, "desc": "调用密钥 Secret"}, {"name": "api_name", "type": "String", "required": true, "desc": "API 接口名称，固定为 t_img"}, {"name": "result_type", "type": "String", "required": false, "desc": "返回格式：json/jsonu/xml/serialize，默认 json"}, {"name": "lang", "type": "String", "required": false, "desc": "接口说明语言：cn/en/ru，默认 cn"}, {"name": "cache", "type": "String", "required": false, "desc": "是否使用缓存：yes/no，默认 yes"} ], "business_params": [ {"name": "imgcode", "type": "String", "required": true, "desc": "图片地址（URL）或图片 Base64 编码（POST 方式）"}, {"name": "to", "type": "String", "required": true, "desc": "目标语言代码，如 en（英语）、jp（日语）、kor（韩语）等"} ] }, { "section_title": "三、返回数据结构详解", "content": "接口返回 JSON 格式数据，顶层为 items 对象，包含源语言、目标语言以及识别翻译内容列表。每个 content 元素对应图片中一行识别到的文字，包含原文、译文、位置坐标等完整信息。", "response_structure": { "items": { "from": "源语言代码（如 zh）", "to": "目标语言代码（如 en）", "content": [ { "src": "识别到的原始文本", "dst": "翻译后的文本", "rect": "文字区域矩形坐标（x y width height）", "lineCount": "文本行数", "points": "文字区域四角坐标点数组 [{x,y}, ...]", "pasteImg": "回填效果图（Base64，部分场景返回）" } ] } }, "response_example": { "items": { "from": "zh", "to": "en", "content": [ { "src": "恭喜发财", "dst": "May you be happy and prosperous", "rect": "209 3 382 78", "lineCount": 1, "points": [ {"x": 209, "y": 3}, {"x": 589, "y": 3}, {"x": 589, "y": 78}, {"x": 209, "y": 78} ], "pasteImg": "" }, { "src": "小财神", "dst": "Little God of Wealth", "rect": "612 641 157 53", "lineCount": 1, "points": [ {"x": 612, "y": 641}, {"x": 768, "y": 641}, {"x": 768, "y": 692}, {"x": 612, "y": 692} ], "pasteImg": "" } ] } } }, { "section_title": "四、支持语言列表", "content": "t_img 接口覆盖主流语种，常用语言代码对照如下：", "language_codes": [ {"language": "中文", "code": "zh"}, {"language": "英语", "code": "en"}, {"language": "日语", "code": "jp"}, {"language": "韩语", "code": "kor"}, {"language": "法语", "code": "fra"}, {"language": "德语", "code": "de"}, {"language": "西班牙语", "code": "spa"}, {"language": "俄语", "code": "ru"}, {"language": "葡萄牙语", "code": "pt"}, {"language": "意大利语", "code": "it"}, {"language": "荷兰语", "code": "nl"}, {"language": "马来语", "code": "may"}, {"language": "印尼语", "code": "id"}, {"language": "土耳其语", "code": "tr"}, {"language": "波兰语", "code": "pl"}, {"language": "瑞典语", "code": "swe"}, {"language": "丹麦语", "code": "dan"}, {"language": "希腊语", "code": "el"}, {"language": "匈牙利语", "code": "hu"}, {"language": "罗马尼亚语", "code": "rom"} ] }, { "section_title": "五、Python 接入实战", "content": "以下是基于 requests 库的 Python 调用示例，涵盖 URL 图片和 Base64 图片两种传入方式，以及返回结果的结构化解析。", "code_example_python": """import requests import base64 from typing import Dict, List, Any class TImgTranslator: """t_img 图片翻译客户端""" def __init__(self, api_key: str, api_secret: str): self.api_key = api_key self.api_secret = api_secret self.base_url = "https://api-gw.onebound.cn/translate/t_img" def translate_by_url(self, img_url: str, target_lang: str = "en") -> Dict[str, Any]: """通过图片 URL 翻译""" params = { "key": self.api_key, "secret": self.api_secret, "api_name": "t_img", "imgcode": img_url, "to": target_lang, "result_type": "json" } resp = requests.get(self.base_url, params=params, timeout=30) resp.raise_for_status() return resp.json() def translate_by_base64(self, image_path: str, target_lang: str = "en") -> Dict[str, Any]: """通过本地图片 Base64 翻译""" with open(image_path, "rb") as f: img_base64 = base64.b64encode(f.read()).decode() payload = { "key": self.api_key, "secret": self.api_secret, "api_name": "t_img", "imgcode": img_base64, "to": target_lang } resp = requests.post(self.base_url, json=payload, timeout=30) resp.raise_for_status() return resp.json() def extract_translations(self, result: Dict[str, Any]) -> List[Dict[str, str]]: """提取翻译文本对列表""" items = result.get("items", {}) content_list = items.get("content", []) return [ { "original": item.get("src", ""), "translated": item.get("dst", ""), "position": item.get("rect", "") } for item in content_list ] # 使用示例 if __name__ == "__main__": translator = TImgTranslator("your_api_key", "your_api_secret") result = translator.translate_by_url( img_url="https://example.com/test.jpg", target_lang="en" ) translations = translator.extract_translations(result) for t in translations: print(f"原文: {t['original']}") print(f"译文: {t['translated']}") print(f"位置: {t['position']}") print("---")""" }, { "section_title": "六、Java 接入示例", "content": "Java 环境下通过原生 URLConnection 调用接口，解析 JSON 返回结果。", "code_example_java": """import java.io.BufferedReader; import java.io.InputStreamReader; import java.net.URL; import java.net.URLConnection; import java.nio.charset.StandardCharsets; import org.json.JSONObject; import org.json.JSONArray; public class TImgExample { private static final String API_URL = "https://api-gw.onebound.cn/translate/t_img/"; public static JSONObject translateImage(String key, String secret, String imgUrl, String targetLang) throws Exception { String url = API_URL + "?key=" + key + "&secret=" + secret + "&api_name=t_img&imgcode=" + imgUrl + "&to=" + targetLang; URL realUrl = new URL(url); URLConnection conn = realUrl.openConnection(); conn.setConnectTimeout(10000); conn.setReadTimeout(30000); try (BufferedReader rd = new BufferedReader( new InputStreamReader(conn.getInputStream(), StandardCharsets.UTF_8))) { StringBuilder sb = new StringBuilder(); int cp; while ((cp = rd.read()) != -1) { sb.append((char) cp); } return new JSONObject(sb.toString()); } } public static void main(String[] args) throws Exception { JSONObject result = translateImage( "your_api_key", "your_api_secret", "https://example.com/test.jpg", "en" ); JSONObject items = result.getJSONObject("items"); JSONArray content = items.getJSONArray("content"); System.out.println("源语言: " + items.getString("from")); System.out.println("目标语言: " + items.getString("to")); System.out.println("识别行数: " + content.length()); for (int i = 0; i < content.length(); i++) { JSONObject item = content.getJSONObject(i); System.out.println("原文: " + item.getString("src")); System.out.println("译文: " + item.getString("dst")); } } }""" }, { "section_title": "七、应用场景", "content": "t_img 接口的应用场景覆盖跨境电商、内容平台、工具类产品等多个领域。", "scenarios": [ { "name": "跨境电商商品图翻译", "desc": "批量翻译商品详情图、海报图中的文字，快速适配多语言站点，提升海外用户转化率" }, { "name": "截图实时翻译工具", "desc": "集成到浏览器插件、桌面应用中，实现选区截图即翻译，适合阅读外文文档、网页时使用" }, { "name": "文档图片翻译", "desc": "扫描件、PDF 转图片后批量翻译，保留版式坐标信息，便于还原排版输出" }, { "name": "APP 拍照翻译", "desc": "移动端调用摄像头拍照后上传翻译，应用于旅游翻译、菜单翻译、路牌翻译等场景" }, { "name": "社交媒体内容本地化", "desc": "自动识别并翻译社交媒体图片中的文字，助力内容多语言分发和运营" } ] }, { "section_title": "八、最佳实践与注意事项", "content": "在生产环境接入 t_img 接口时，建议从异常处理、缓存策略、图片预处理等方面做好工程化保障。", "best_practices": [ "图片大小控制在 10MB 以内，过大图片建议先压缩再上传，可显著提升响应速度", "优先使用图片 URL 方式传入，减少请求体大小；本地图片再走 Base64 路径", "开启 cache=yes 利用缓存机制，相同图片重复翻译可大幅降低调用成本和延迟", "对返回的 content 列表做空判断，图片中无文字时 content 可能为空数组", "rect 字段格式为「x y width height」，points 为四角坐标，可根据需要选择使用", "生产环境建议配置超时重试机制，网络波动时自动重试 2-3 次保障成功率", "批量翻译场景建议控制并发数，避免触发频率限制，可结合队列削峰处理", "敏感内容翻译需结合内容安全接口做二次审核，确保合规" ]