如何ocr识别图片中是否有logo？4种思路解决方案分析

谁践踏了优雅 2022-10-31 04:09 88阅读 0赞

识别图片中是否有logo，实现方案主要有如下4种方案：

**目录**

1.通过logo库来判断图片中是否有logo库中的logo

2.识别图片中的文字，只要有文字，就假设有logo，再进行二次校验

3.识别图片中的文字，通过logo文字库，来判断是否是logo

4.直接识别logo，但需要通过大量带logo图片进行训练，来实现识别的准确性

--------------------

# 1.通过logo库来判断图片中是否有logo库中的logo #

百度logo识别接口

接口地址参考：[https://cloud.baidu.com/doc/IMAGERECOGNITION/s/Ok3bcxc59][https_cloud.baidu.com_doc_IMAGERECOGNITION_s_Ok3bcxc59]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2pscV9kaWxpZ2VuY2U_size_16_color_FFFFFF_t_70][]

阿里logo识别接口

阿里logo识别地址参考：[https://help.aliyun.com/knowledge\_detail/155012.html][https_help.aliyun.com_knowledge_detail_155012.html]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2pscV9kaWxpZ2VuY2U_size_16_color_FFFFFF_t_70 1][]

缺点：这种方式依赖logo库

# 2.识别图片中的文字，只要有文字，就假设有logo，再进行二次校验 #

识别图片中的文字，目前不管自己用Java/Python写代码识别图片中文字，还是调用第三方接口，技术都已比较成熟。

下面以其中一种方式举例说明：

直接上代码：

public static String generalBasic(String filePath,String accessToken) {
            // 请求url
            String url = "https://aip.baidubce.com/rest/2.0/ocr/v1/general_basic";
            try {
                byte[] imgData = FileUtil.readFileByBytes(filePath);
                String imgStr = Base64Util.encode(imgData);
                String imgParam = URLEncoder.encode(imgStr, "UTF-8");
    
                String param = "image=" + imgParam;
    
                String result = HttpUtil.post(url, accessToken, param);
                return getWords(result);
            } catch (Exception e) {
                e.printStackTrace();
            }
            return null;
        }
        public static void main(String[] args) {
            String filePath = "D:\\1.jpg";
            String accessToken = "百度api的token";
           String word=generalBasic(filePath,accessToken);
            System.out.println(word);
        }

其他Java类如下：

编码处理类：

package com.example.demo.ocr.common;
    
    /**
     * Base64 工具类
     */
    public class Base64Util {
        private static final char last2byte = (char) Integer.parseInt("00000011", 2);
        private static final char last4byte = (char) Integer.parseInt("00001111", 2);
        private static final char last6byte = (char) Integer.parseInt("00111111", 2);
        private static final char lead6byte = (char) Integer.parseInt("11111100", 2);
        private static final char lead4byte = (char) Integer.parseInt("11110000", 2);
        private static final char lead2byte = (char) Integer.parseInt("11000000", 2);
        private static final char[] encodeTable = new char[]{'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', '+', '/'};
    
        public Base64Util() {
        }
    
        public static String encode(byte[] from) {
            StringBuilder to = new StringBuilder((int) ((double) from.length * 1.34D) + 3);
            int num = 0;
            char currentByte = 0;
    
            int i;
            for (i = 0; i < from.length; ++i) {
                for (num %= 8; num < 8; num += 6) {
                    switch (num) {
                        case 0:
                            currentByte = (char) (from[i] & lead6byte);
                            currentByte = (char) (currentByte >>> 2);
                        case 1:
                        case 3:
                        case 5:
                        default:
                            break;
                        case 2:
                            currentByte = (char) (from[i] & last6byte);
                            break;
                        case 4:
                            currentByte = (char) (from[i] & last4byte);
                            currentByte = (char) (currentByte << 2);
                            if (i + 1 < from.length) {
                                currentByte = (char) (currentByte | (from[i + 1] & lead2byte) >>> 6);
                            }
                            break;
                        case 6:
                            currentByte = (char) (from[i] & last2byte);
                            currentByte = (char) (currentByte << 4);
                            if (i + 1 < from.length) {
                                currentByte = (char) (currentByte | (from[i + 1] & lead4byte) >>> 4);
                            }
                    }
    
                    to.append(encodeTable[currentByte]);
                }
            }
    
            if (to.length() % 4 != 0) {
                for (i = 4 - to.length() % 4; i > 0; --i) {
                    to.append("=");
                }
            }
    
            return to.toString();
        }
    }

读图片类：

/**
         * 根据文件路径读取byte[] 数组
         */
        public static byte[] readFileByBytes(String filePath) throws IOException {
            File file = new File(filePath);
            if (!file.exists()) {
                throw new FileNotFoundException(filePath);
            } else {
                ByteArrayOutputStream bos = new ByteArrayOutputStream((int) file.length());
                BufferedInputStream in = null;
    
                try {
                    in = new BufferedInputStream(new FileInputStream(file));
                    short bufSize = 1024;
                    byte[] buffer = new byte[bufSize];
                    int len1;
                    while (-1 != (len1 = in.read(buffer, 0, bufSize))) {
                        bos.write(buffer, 0, len1);
                    }
    
                    byte[] var7 = bos.toByteArray();
                    return var7;
                } finally {
                    try {
                        if (in != null) {
                            in.close();
                        }
                    } catch (IOException var14) {
                        var14.printStackTrace();
                    }
    
                    bos.close();
                }
            }
        }

HttpUtil请求类：

/**
     * http 工具类
     */
    public class HttpUtil {
    
        public static String post(String requestUrl, String accessToken, String params)
                throws Exception {
            String contentType = "application/x-www-form-urlencoded";
            return HttpUtil.post(requestUrl, accessToken, contentType, params);
        }
    
        public static String post(String requestUrl, String accessToken, String contentType, String params)
                throws Exception {
            String encoding = "UTF-8";
            if (requestUrl.contains("nlp")) {
                encoding = "GBK";
            }
            return HttpUtil.post(requestUrl, accessToken, contentType, params, encoding);
        }
    
        public static String post(String requestUrl, String accessToken, String contentType, String params, String encoding)
                throws Exception {
            String url = requestUrl + "?access_token=" + accessToken;
            return HttpUtil.postGeneralUrl(url, contentType, params, encoding);
        }
    
        public static String postGeneralUrl(String generalUrl, String contentType, String params, String encoding)
                throws Exception {
            URL url = new URL(generalUrl);
            // 打开和URL之间的连接
            HttpURLConnection connection = (HttpURLConnection) url.openConnection();
            connection.setRequestMethod("POST");
            // 设置通用的请求属性
            connection.setRequestProperty("Content-Type", contentType);
            connection.setRequestProperty("Connection", "Keep-Alive");
            connection.setUseCaches(false);
            connection.setDoOutput(true);
            connection.setDoInput(true);
    
            // 得到请求的输出流对象
            DataOutputStream out = new DataOutputStream(connection.getOutputStream());
            out.write(params.getBytes(encoding));
            out.flush();
            out.close();
    
            // 建立实际的连接
            connection.connect();
            // 获取所有响应头字段
            Map<String, List<String>> headers = connection.getHeaderFields();
            // 遍历所有的响应头字段
    //        for (String key : headers.keySet()) {
    //            System.err.println(key + "--->" + headers.get(key));
    //        }
            // 定义 BufferedReader输入流来读取URL的响应
            BufferedReader in = null;
            in = new BufferedReader(
                    new InputStreamReader(connection.getInputStream(), encoding));
            String result = "";
            String getLine;
            while ((getLine = in.readLine()) != null) {
                result += getLine;
            }
            in.close();
    //        System.err.println("result:" + result);
            return result;
        }
    }

获取百度token：

import org.json.JSONObject;
    
    import java.io.BufferedReader;
    import java.io.InputStreamReader;
    import java.net.HttpURLConnection;
    import java.net.URL;
    import java.util.List;
    import java.util.Map;
    
    /**
     * 获取token类
     */
    public class AuthService {
        public static void main(String[] args){
            String token=getAuth();
            System.out.println(token);
        }
        /**
         * 获取权限token
         * @return 返回示例：
         * {
         * "access_token": "24.460da4889caad24cccdb1fea17221975.2592000.1491995545.282335-1234567",
         * "expires_in": 2592000
         * }
         */
        public static String getAuth() {
            // 官网获取的 API Key 更新为你注册的
            String clientId = "您的API Key";
            // 官网获取的 Secret Key 更新为你注册的
            String clientSecret = "您的Secret Key";
            return getAuth(clientId, clientSecret);
        }
    
        /**
         * 获取API访问token
         * 该token有一定的有效期，需要自行管理，当失效时需重新获取.
         * @param ak - 百度云官网获取的 API Key
         * @param sk - 百度云官网获取的 Securet Key
         * @return assess_token 示例：
         * "24.460da4889caad24cccdb1fea17221975.2592000.1491995545.282335-1234567"
         */
        public static String getAuth(String ak, String sk) {
            // 获取token地址
            String authHost = "https://aip.baidubce.com/oauth/2.0/token?";
            String getAccessTokenUrl = authHost
                    // 1. grant_type为固定参数
                    + "grant_type=client_credentials"
                    // 2. 官网获取的 API Key
                    + "&client_id=" + ak
                    // 3. 官网获取的 Secret Key
                    + "&client_secret=" + sk;
            try {
                URL realUrl = new URL(getAccessTokenUrl);
                // 打开和URL之间的连接
                HttpURLConnection connection = (HttpURLConnection) realUrl.openConnection();
                connection.setRequestMethod("GET");
                connection.connect();
                // 获取所有响应头字段
                Map<String, List<String>> map = connection.getHeaderFields();
                // 遍历所有的响应头字段
                for (String key : map.keySet()) {
                    System.err.println(key + "--->" + map.get(key));
                }
                // 定义 BufferedReader输入流来读取URL的响应
                BufferedReader in = new BufferedReader(new InputStreamReader(connection.getInputStream()));
                String result = "";
                String line;
                while ((line = in.readLine()) != null) {
                    result += line;
                }
                /**
                 * 返回结果示例
                 */
                System.err.println("result:" + result);
                JSONObject jsonObject = new JSONObject(result);
                String access_token = jsonObject.getString("access_token");
                return access_token;
            } catch (Exception e) {
                System.err.printf("获取token失败！");
                e.printStackTrace(System.err);
            }
            return null;
        }
    
    }

百度接口调用及简单二次开发：

private static String getWords(String json) {
            JSONObject obj=new JSONObject(json);
            Integer words_result_num=obj.getInt("words_result_num");
            if(words_result_num>0){
                JSONArray list=obj.getJSONArray("words_result");
                String words="";
                for(int i=0;i<list.length();i++){
                    JSONObject obj2=list.getJSONObject(i);
                    words=words+("".equals(words)?"":",")+obj2.get("words");
                    //System.out.println(list.get(i));
                    //System.out.println(obj2.get("words"));
                }
                return words;
            }
            return "";
        }
        public static String generalBasic(String filePath,String accessToken) {
            // 请求url
            String url = "https://aip.baidubce.com/rest/2.0/ocr/v1/general_basic";
            try {
                byte[] imgData = FileUtil.readFileByBytes(filePath);
                String imgStr = Base64Util.encode(imgData);
                String imgParam = URLEncoder.encode(imgStr, "UTF-8");
    
                String param = "image=" + imgParam;
    
                String result = HttpUtil.post(url, accessToken, param);
                return getWords(result);
            } catch (Exception e) {
                e.printStackTrace();
            }
            return null;
        }

下面拿2张图，识别出的logo举例：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2pscV9kaWxpZ2VuY2U_size_16_color_FFFFFF_t_70 2][]

这张图，识别出“hzfwjgs cn alibaba. com”，如下字符串：

![20210222221914655.png][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2pscV9kaWxpZ2VuY2U_size_16_color_FFFFFF_t_70 3][]

这张图，识别出“分南宁市第一中学”，如下字符串：

![20210222221753832.png][]

完整代码地址 [http://www.zrscsoft.com/sitepic/12130.html][http_www.zrscsoft.com_sitepic_12130.html]

# 3.识别图片中的文字，通过logo文字库，来判断是否是logo #

这种方式，是在方案2基础上继续延续的方案。

主要分三步实现：

1）、识别图片中的文字

详细见方案2，这里略。

2）、建立logo文字库

建立logo文字库，细节就不详细说。举个例子吧。

logo文字库如“www”，“com”，“网”，“媒体”，“新闻”等等。

3）、使用logo文字库来判断图片是否带logo

根据图片中的文字，检索logo文字库中的文字，例如图片中带“www”，“com”，“网”，“媒体”，“新闻”就判断为是带logo的图片

完整项目地址：[http://www.zrscsoft.com/sitepic/12130.html][http_www.zrscsoft.com_sitepic_12130.html]

# 4.直接识别logo，但需要通过大量带logo图片进行训练，来实现识别的准确性 #

这种方式，目前还没有找到实现方法。如果有实现，欢迎在评论区留言。

[https_cloud.baidu.com_doc_IMAGERECOGNITION_s_Ok3bcxc59]: https://cloud.baidu.com/doc/IMAGERECOGNITION/s/Ok3bcxc59
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2pscV9kaWxpZ2VuY2U_size_16_color_FFFFFF_t_70]: /images/20221024/d2bd0ce9a7f94d1c86e9bf116c0d4449.png
[https_help.aliyun.com_knowledge_detail_155012.html]: https://help.aliyun.com/knowledge_detail/155012.html
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2pscV9kaWxpZ2VuY2U_size_16_color_FFFFFF_t_70 1]: /images/20221024/c9d6a4cb1972489c827189295d81f26f.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2pscV9kaWxpZ2VuY2U_size_16_color_FFFFFF_t_70 2]: /images/20221024/c571d719c87c455ab1cf523e7e3ea504.png
[20210222221914655.png]: /images/20221024/fce2b331e8aa4981a0af1034bb3d079a.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2pscV9kaWxpZ2VuY2U_size_16_color_FFFFFF_t_70 3]: /images/20221024/fb7af95ad8294cff9e1d47cdc821dfb1.png
[20210222221753832.png]: /images/20221024/cac4ab791fca463c883e9d1d8cc8ca03.png
[http_www.zrscsoft.com_sitepic_12130.html]: http://www.zrscsoft.com/sitepic/12130.html