目前通用物体和场景识别支持超过 10 万类常见物体和场景识别,接口返回图片内 1 个或多个物体的名称,并可获取百科信息。适用于图像或视频内容分析、拍照识图等业务场景。

功能介绍

1. 识别物体或场景名称

支持识别动物、植物、商品、建筑、风景、动漫、食材、公众人物等 10 万个常见物体及场景,接口返回大类及细分类的名称结果。

2. 获取百科信息

支持获取图片识别结果对应的百科信息,接口返回百科词条 URL、图片和摘要描述,可选择是否需要返回百科信息。

应用场景

  • 图片内容分析与推荐:对用户浏览的图片或观看的视频内容进行识别,根据识别结果给出相关内容推荐或广告展示。广泛应用于新闻资讯类、视频类 app 等内容平台中
  • 拍照识图:根据用户拍摄照片,识别图片中物体名称及百科信息,提高用户交互体验,广泛应用于智能手机厂商、拍照识图及科普类 app 中
  • 拍照闯关趣味营销:设计线上营销活动,根据用户拍摄照片,自动识别图片中物体信息是否符合活动要求,提升用户交互体验,减少人工审核成本

接口描述

该请求用于通用物体及场景识别,即对于输入的一张图片(可正常解码,且长宽比适宜),输出图片中的多个物体及场景标签。

请求说明

  • HTTP 方法: POST
  • 请求 URL: https://aip.baidubce.com/rest/2.0/image-classify/v2/advanced_general
  • URL参数: access_token
  • Header 参数: Content-Type = application/x-www-form-urlencoded
  • Body 参数:见下表
参数 是否必选 类型 可选值范围 说明
image true string - 图像数据,base64 编码,要求 base64 编码后大小不超过 4M,最短边至少 15px,最长边最大4096px,支持 jpg/png/bmp 格式。注意:图片需要 base64 编码、去掉编码头(data:image/jpg;base64,)后,再进行 urlencode
baike_num integer 0 返回百科信息的结果数,默认不返回

返回说明

返回参数如下表:

字段 是否必选 类型 说明
log_id uint64 唯一的 log id,用于问题定位
result_num unit32 返回结果数目,及 result 数组中的元素个数
result arrry(object) 标签结果数组
+keyword string 图片中的物体或场景名称
+score float 置信度,0 - 1
+root string 识别结果的上层标签,有部分钱币、动漫、烟酒等 tag 无上层标签
+baike_info object 对应识别结果的百科词条名称
++baike_url string 对应识别结果百度百科页面链接
++image_url string 对应识别结果百科图片链接
++description string 对应识别结果百科内容描述

返回示例如下:

{
"log_id": "4269290077301074002",
"result_num": 5,
"result": [
{
"score": 0.203018,
"root": "植物-树",
"baike_info": {
"baike_url": "http://baike.baidu.com/item/%E6%A0%91/2699484",
"image_url": "http://imgsrc.baidu.com/baike/pic/item/6159252dd42a2834218a2c2154b5c9ea15cebfef.jpg",
"description": "树状图是一种数据结构,它是由n(n>=1)个有限结点组成一个具有层次关系的集合。把它叫做“树”是因为它看起来像一棵倒挂的树,也就是说它是根朝上,而叶朝下的。它具有以下的特点:每个结点有零个或多个子结点;没有父结点的结点称为根结点;每一个非根结点有且只有一个父结点;除了根结点外,每个子结点可以分为多个不相交的子树;"
},
"keyword": "树"
},
{
"score": 0.15084,
"root": "自然风景-其他",
"keyword": "风景"
},
{
"score": 0.100728,
"root": "自然风景-天空",
"keyword": "天空"
},
{
"score": 0.052704,
"root": "商品-工艺品",
"keyword": "工艺品"
},
{
"score": 0.003729,
"root": "商品-工艺品",
"keyword": "佛像"
}
]
}

C++ 代码实现调用

这里假设已经将环境配置好了,环境配置的文章可以参考 Windows 下使用 Vcpkg 配置百度 AI 图像识别 C++开发环境(VS2017)

为了方便,首先根据返回参数定义了一个结构体,该结构体包括了返回参数中的参数,如下:

struct GeneralInfo {
std::string keyword;
float score;
std::string root;
std::string baikeurl;
std::string imageurl;
std::string baikedesc;

void print() {
std::cout << std::setw(30) << std::setfill('-') << '\n';
std::cout << "keyword: " << keyword << "\n";
std::cout << "score: " << std::fixed << std::setprecision(4) << score << "\n";
std::cout << "root: " << root << "\n";
if (baikeurl != "null")
std::cout << "baikeurl: " << baikeurl << "\n";
if (imageurl != "null")
std::cout << "imageurl: " << imageurl << "\n";
if (baikedesc != "null")
std::cout << "baikedesc: " << baikedesc << "\n";
}
};

然后定义了一个类来调用接口并获取结果

class General
{
public:
General();
~General();

Json::Value request(std::string imgBase64, std::map<std::string, std::string>& options);

int getResultNum();

// get all return results
void getAllResult(std::vector<GeneralInfo>& results);

// only get first result
void getResult(GeneralInfo& result);


private:
Json::Value obj_;
std::string url_;
uint32_t result_num_;
// file to save token key
std::string filename_;
};

类中的私有成员 obj_ 表示返回结果对应的 json 对象。url_ 表示请求的 url,result_num_ 表示返回结果数,filename_ 表示用于存储 access token 的文件的文件名。

request 函数输入请求图像的 base64 编码以及请求参数,返回一个 json 对象,json 对象中包含请求的结果。

getAllResult 将所有返回结果存放到一个 vector 中。getResult 只返回第一条结果,也就是分数最高的结果。


完整代码如下

General.h 代码如下:

#pragma once
#include "util.h"

struct GeneralInfo {
std::string keyword;
float score;
std::string root;
std::string baikeurl;
std::string imageurl;
std::string baikedesc;

void print() {
std::cout << std::setw(30) << std::setfill('-') << '\n';
std::cout << "keyword: " << keyword << "\n";
std::cout << "score: " << std::fixed << std::setprecision(4) << score << "\n";
std::cout << "root: " << root << "\n";
if (baikeurl != "null")
std::cout << "baikeurl: " << baikeurl << "\n";
if (imageurl != "null")
std::cout << "imageurl: " << imageurl << "\n";
if (baikedesc != "null")
std::cout << "baikedesc: " << baikedesc << "\n";
}
};

class General
{
public:
General();
~General();

Json::Value request(std::string imgBase64, std::map<std::string, std::string>& options);

int getResultNum();

// get all return results
void getAllResult(std::vector<GeneralInfo>& results);

// only get first result
void getResult(GeneralInfo& result);


private:
Json::Value obj_;
std::string url_;
uint32_t result_num_;
// file to save token key
std::string filename_;
};

void generalTest();

General.cpp 代码如下:

#include "General.h"

General::General()
{
filename_ = "tokenKey";
url_ = "https://aip.baidubce.com/rest/2.0/image-classify/v2/advanced_general";
}


General::~General()
{
}

Json::Value General::request(std::string imgBase64, std::map<std::string, std::string>& options)
{
std::string response;
Json::Value obj;
std::string token;

// 1. get HTTP post body
std::string body;
mergeHttpPostBody(body, imgBase64, options);

// 2. get HTTP url with access token
std::string url = url_;
getHttpPostUrl(url, filename_, token);

// 3. post request, response store the result
int status_code = httpPostRequest(url, body, response);
if (status_code != CURLcode::CURLE_OK) {
obj["curl_error_code"] = status_code;
obj_ = obj;
return obj; // TODO: maybe should exit
}

// 4. make string to json object
generateJson(response, obj);

// if access token is invalid or expired, we will get a new one
if (obj["error_code"].asInt() == 110 || obj["error_code"].asInt() == 111) {
token = getTokenKey();
writeFile(filename_, token);
return request(imgBase64, options);
}

obj_ = obj;

checkErrorWithExit(obj);

result_num_ = getResultNum();

return obj;
}

int General::getResultNum()
{
return obj_["result_num"].asInt();
}

void General::getAllResult(std::vector<GeneralInfo>& results)
{
results.reserve(result_num_);
GeneralInfo tmp;

for (uint32_t i = 0; i < result_num_; ++i) {
tmp.keyword = UTF8ToGB(obj_["result"][i]["keyword"].asString().c_str());
tmp.score = obj_["result"][i]["score"].asFloat();
tmp.root = UTF8ToGB(obj_["result"][i]["root"].asString().c_str());
tmp.baikeurl = obj_["result"][i]["baike_info"].get("baike_url", "null").asString();
tmp.imageurl = obj_["result"][i]["baike_info"].get("image_url", "null").asString();
tmp.baikedesc = UTF8ToGB(obj_["result"][i]["baike_info"].get("description", "null").asString().c_str());
results.push_back(tmp);
}
}

void General::getResult(GeneralInfo& result)
{
result.keyword = UTF8ToGB(obj_["result"][0]["keyword"].asString().c_str());
result.score = obj_["result"][0]["score"].asFloat();
result.root = UTF8ToGB(obj_["result"][0]["root"].asString().c_str());
result.baikeurl = obj_["result"][0]["baike_info"].get("baike_url", "null").asString();
result.imageurl = obj_["result"][0]["baike_info"].get("image_url", "null").asString();
result.baikedesc = UTF8ToGB(obj_["result"][0]["baike_info"].get("description", "null").asString().c_str());
}


void generalTest()
{

std::cout << "size: " << sizeof(GeneralInfo) << "\n";

// read image and encode to base64
std::string imgFile = "./images/cat.jpg";
std::string imgBase64;
imgToBase64(imgFile, imgBase64);

// set options
std::map<std::string, std::string> options;
options["baike_num"] = "5";


Json::Value obj;
General generalObj;
obj = generalObj.request(imgBase64, options);

GeneralInfo result;
generalObj.getResult(result);
result.print();

std::vector<GeneralInfo> results;
generalObj.getAllResult(results);

for (auto & vec : results) {
vec.print();
}
}

util.h 代码如下:

#pragma once
#include <curl/curl.h>
#include <string>
#include <map>
#include <iostream>
#include <fstream>
#include <json/json.h>
#include <iomanip>
#include <opencv2/opencv.hpp>
#include "base64.h"

size_t writeCallback(void *ptr, size_t size, size_t nmemb, void *userdata);
std::string getTokenKey();
int readImageFile(const char *filename, std::string& out);
void imgToBase64(std::string &imgFile, std::string &imgBase64);
int writeFile(const std::string & fileString, const std::string &str);
int readFile(const std::string & fileString, std::string &str);
std::string UTF8ToGB(const char* str);
void mergeHttpPostBody(std::string &body, std::string imgBase64, std::map<std::string, std::string>& options);
void getHttpPostUrl(std::string &url, std::string &filename, std::string &token);
int httpPostRequest(std::string &url, std::string &body, std::string &response);
void generateJson(std::string &response, Json::Value &obj);
void checkErrorWithExit(Json::Value &obj);

util.cpp 代码如下:

#include "util.h"

// callback function for curl
size_t writeCallback(void *ptr, size_t size, size_t nmemb, void *userdata)
{
std::string *str = dynamic_cast<std::string *>((std::string *)userdata);
str->append((char *)ptr, size * nmemb);
return size * nmemb;
}

// get access token from server by get method
std::string getTokenKey() {
std::string url = "https://aip.baidubce.com/oauth/2.0/token";
std::string apikey = "FP6E0dqEnGTPAXMYpUDXW7tb";
std::string secritkey = "MoNhkCh3VhtXpMft7Xoj4hQy4iHVbNZu";
std::map<std::string, std::string> params;
std::string response;

params["grant_type"] = "client_credentials";
params["client_id"] = apikey;
params["client_secret"] = secritkey;

// append url with parameters
for (auto it = params.begin(); it != params.end(); ++it) {
url += (it == params.begin() ? "?" : "&") + it->first + "=" + it->second;
}

CURL *curl = curl_easy_init();

struct curl_slist * slist = NULL;

curl_easy_setopt(curl, CURLOPT_URL, url.c_str());
curl_easy_setopt(curl, CURLOPT_HTTPHEADER, slist);
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, writeCallback); // set callback function
curl_easy_setopt(curl, CURLOPT_WRITEDATA, &response); // set var to receive return info from callback function
curl_easy_setopt(curl, CURLOPT_NOSIGNAL, true);
curl_easy_setopt(curl, CURLOPT_SSL_VERIFYPEER, false);
curl_easy_setopt(curl, CURLOPT_SSL_VERIFYHOST, false);
curl_easy_setopt(curl, CURLOPT_VERBOSE, false);

int status_code = curl_easy_perform(curl);

curl_easy_cleanup(curl);
curl_slist_free_all(slist);

Json::Value obj;
if (status_code != CURLcode::CURLE_OK) {
obj["curl_error_code"] = status_code;
return obj.toStyledString();
}

// parse json string
JSONCPP_STRING error;
Json::CharReaderBuilder builder;
const std::unique_ptr<Json::CharReader> reader(builder.newCharReader());
reader->parse(response.data(), response.data() + response.size(), &obj, &error);
std::string access_token = obj["access_token"].asString();

return access_token;
}

// read image file [https://stackoverflow.com/questions/9612121/how-to-read-image-files-and-store-it-in-memorystdstring-in-c]
int readImageFile(const char *filename, std::string& out) {
std::ifstream in(filename, std::ios::in | std::ios::binary);
if (in) {
std::ostringstream oss;
oss << in.rdbuf();
out.assign(oss.str());
return 0;
}
else {
std::cerr << "Can't open image!" << std::endl;
return -1;
}
}

void imgToBase64(std::string &imgFile, std::string &imgBase64) {
// read image and encode to base64
std::string out;
readImageFile(imgFile.c_str(), out);
imgBase64 = base64_encode(out.c_str(), (int)out.size());
}

int writeFile(const std::string & fileString, const std::string &str) {
std::ofstream out(fileString, std::ios::binary);
if (out.is_open()) {
out << str;
out.close();
}

return 0;
}

int readFile(const std::string & fileString, std::string &str) {
std::ifstream in(fileString);
if (!in.is_open()) {
str = "";
return -1;
}

char buffer[256];
while (!in.eof()) {
in.getline(buffer, sizeof(buffer));
}

str = buffer;
return 0;
}


std::string UTF8ToGB(const char* str)
{
std::string result;
WCHAR *strSrc;
LPSTR szRes;

int i = MultiByteToWideChar(CP_UTF8, 0, str, -1, NULL, 0);
strSrc = new WCHAR[i + 1];
MultiByteToWideChar(CP_UTF8, 0, str, -1, strSrc, i);

i = WideCharToMultiByte(CP_ACP, 0, strSrc, -1, NULL, 0, NULL, NULL);
szRes = new CHAR[i + 1];
WideCharToMultiByte(CP_ACP, 0, strSrc, -1, szRes, i, NULL, NULL);

result = szRes;
delete[]strSrc;
delete[]szRes;
return result;
}

void mergeHttpPostBody(std::string &body, std::string imgBase64, std::map<std::string, std::string>& options) {
body = "image=" + std::string(curl_escape(imgBase64.c_str(), int(imgBase64.size()))) + "&";
// append body with options
for (auto it = options.begin(); it != options.end(); ++it) {
body += std::string(curl_escape(it->first.c_str(), (int)it->first.size()))
+ "=" + std::string(curl_escape(it->second.c_str(), (int)it->second.size())) + "&";
}
}

// first get token from file (if exist), then add token to url
void getHttpPostUrl(std::string &url, std::string &filename, std::string &token) {
// if token file is not exist, a new one should be create
if (readFile(filename, token) < 0) {
token = getTokenKey();
writeFile(filename, token);
}

url = url + "?access_token=" + token;
}

int httpPostRequest(std::string &url, std::string &body, std::string &response) {
struct curl_slist * slist = NULL;

CURL *curl = curl_easy_init();

// set headers, actually this is used by default
std::string headers = "Content-Type:application/x-www-form-urlencoded";
slist = curl_slist_append(slist, headers.c_str());

curl_easy_setopt(curl, CURLOPT_URL, url.c_str());
curl_easy_setopt(curl, CURLOPT_HTTPHEADER, slist);
curl_easy_setopt(curl, CURLOPT_POST, true);
curl_easy_setopt(curl, CURLOPT_POSTFIELDS, body.data());
curl_easy_setopt(curl, CURLOPT_POSTFIELDSIZE, body.size());
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, writeCallback);
curl_easy_setopt(curl, CURLOPT_WRITEDATA, (void *)&response);
curl_easy_setopt(curl, CURLOPT_NOSIGNAL, true);
curl_easy_setopt(curl, CURLOPT_SSL_VERIFYPEER, false);
curl_easy_setopt(curl, CURLOPT_SSL_VERIFYHOST, false);
curl_easy_setopt(curl, CURLOPT_VERBOSE, false);

int status_code = curl_easy_perform(curl);

curl_easy_cleanup(curl);
curl_slist_free_all(slist);

return status_code;
}

void generateJson(std::string &response, Json::Value &obj) {
// parse json string
JSONCPP_STRING error;
Json::CharReaderBuilder builder;
const std::unique_ptr<Json::CharReader> reader(builder.newCharReader());
reader->parse(response.data(), response.data() + response.size(), &obj, &error);
}

void checkErrorWithExit(Json::Value &obj)
{
if (obj.get("error_code", "null")) {
std::cerr << obj.get("error_code", "null") << " : " << obj.get("error_msg", "null") << std::endl;
system("pause");
exit(EXIT_FAILURE);
}
}

运行结果

测试图像