對於類似以下簡單的驗證碼的識別方案：
1、
2
3
4、
1、建庫：切割驗證碼為單個字符，人工標記，比如：A。
2、識別：給一個驗證碼：切割為單個字符，在庫中查詢識別。
/***
* author:chzeze
* 識別驗證碼並返回
* train_path 驗證碼字母圖庫位置
* 驗證碼圖片緩存位置：Configuration.getProperties("web_save_path")+"/captcha.jpg"
*/
public class AmGetCaptchaTest {
private static Logger logger = Logger.getLogger(AmGetCaptchaTest.class);
private static String train_path = "/data/sata/share_sata/AmazonCrawl/amazonWeb/captcha";
private static Map<BufferedImage, String> trainMap = null;
private static int index = 0;
private static int imgnum = 0;
private static MultiThreadedHttpConnectionManager httpConnectionManager = new MultiThreadedHttpConnectionManager();
private static HttpClient client = new HttpClient(httpConnectionManager);
/* static {
//每主機最大連接數和總共最大連接數，通過hosfConfiguration設置host來區分每個主機
client.getHttpConnectionManager().getParams().setDefaultMaxConnectionsPerHost(8);
client.getHttpConnectionManager().getParams().setMaxTotalConnections(48);
client.getHttpConnectionManager().getParams().setConnectionTimeout(10000);
client.getHttpConnectionManager().getParams().setSoTimeout(10000);
client.getHttpConnectionManager().getParams().setTcpNoDelay(true);
client.getHttpConnectionManager().getParams().setLinger(1000);
//失敗的情況下會進行3次嘗試,成功之后不會再嘗試
client.getHttpConnectionManager().getParams().setParameter(HttpMethodParams.RETRY_HANDLER, new DefaultHttpMethodRetryHandler());
}*/
public static int isBlack(int colorInt) {
Color color = new Color(colorInt);
if (color.getRed() + color.getGreen() + color.getBlue() <= 100) {
return 1;
}
return 0;
}
public static int isWhite(int colorInt) {
Color color = new Color(colorInt);
if (color.getRed() + color.getGreen() + color.getBlue() > 600) {
return 1;
}
return 0;
}
public static BufferedImage removeBackgroud(String picFile)
throws Exception {
BufferedImage img = ImageIO.read(new File(picFile));
img = img.getSubimage(1, 1, img.getWidth() - 2, img.getHeight() - 2);
int width = img.getWidth();
int height = img.getHeight();
double subWidth = width / 5.0;
for (int i = 0; i < 5; i++) {
Map<Integer, Integer> map = new HashMap<Integer, Integer>();
for (int x = (int) (1 + i * subWidth); x < (i + 1) * subWidth
&& x < width - 1; ++x) {
for (int y = 0; y < height; ++y) {
if (isWhite(img.getRGB(x, y)) == 1)
continue;
if (map.containsKey(img.getRGB(x, y))) {
map.put(img.getRGB(x, y), map.get(img.getRGB(x, y)) + 1);
} else {
map.put(img.getRGB(x, y), 1);
}
}
}
int max = 0;
int colorMax = 0;
for (Integer color : map.keySet()) {
if (max < map.get(color)) {
max = map.get(color);
colorMax = color;
}
}
for (int x = (int) (1 + i * subWidth); x < (i + 1) * subWidth
&& x < width - 1; ++x) {
for (int y = 0; y < height; ++y) {
if (img.getRGB(x, y) != colorMax) {
img.setRGB(x, y, Color.WHITE.getRGB());
} else {
img.setRGB(x, y, Color.BLACK.getRGB());
}
}
}
}
return img;
}
public static BufferedImage removeBlank(BufferedImage img) throws Exception {
int width = img.getWidth();
int height = img.getHeight();
int start = 0;
int end = 0;
Label1: for (int y = 0; y < height; ++y) {
for (int x = 0; x < width; ++x) {
if (isBlack(img.getRGB(x, y)) == 1) {
start = y;
break Label1;
}
}
}
Label2: for (int y = height - 1; y >= 0; --y) {
for (int x = 0; x < width; ++x) {
if (isBlack(img.getRGB(x, y)) == 1) {
end = y;
break Label2;
}
}
}
return img.getSubimage(0, start, width, end - start + 1);
}
public static List<BufferedImage> splitImage(BufferedImage img)
throws Exception {
List<BufferedImage> subImgs = new ArrayList<BufferedImage>();
int width = img.getWidth();
int height = img.getHeight();
List<Integer> weightlist = new ArrayList<Integer>();
for (int x = 0; x < width; ++x) {
int count = 0;
for (int y = 0; y < height; ++y) {
if (isBlack(img.getRGB(x, y)) == 1) {
count++;
}
}
weightlist.add(count);
}
for (int i = 0; i < weightlist.size();i++) {
int length = 0;
while (i < weightlist.size() && weightlist.get(i) > 0) {
i++;
length++;
}
if (length > 2) {
subImgs.add(removeBlank(img.getSubimage(i - length, 0,
length, height)));
}
}
return subImgs;
}
public static Map<BufferedImage, String> loadTrainData() throws Exception {
if (trainMap == null) {
Map<BufferedImage, String> map = new HashMap<BufferedImage, String>();
File dir = new File(train_path);
File[] files = dir.listFiles();
for (File file : files) {
map.put(ImageIO.read(file), file.getName().charAt(0) + "");
}
trainMap = map;
}
return trainMap;
}
public static String getSingleCharOcr(BufferedImage img,
Map<BufferedImage, String> map) {
String result = "#";
int width = img.getWidth();
int height = img.getHeight();
int min = width * height;
for (BufferedImage bi : map.keySet()) {
int count = 0;
if (Math.abs(bi.getWidth()-width) > 2)
continue;
int widthmin = width < bi.getWidth() ? width : bi.getWidth();
int heightmin = height < bi.getHeight() ? height : bi.getHeight();
Label1: for (int x = 0; x < widthmin; ++x) {
for (int y = 0; y < heightmin; ++y) {
if (isBlack(img.getRGB(x, y)) != isBlack(bi.getRGB(x, y))) {
count++;
if (count >= min)
break Label1;
}
}
}
if (count < min) {
min = count;
result = map.get(bi);
}
}
return result;
}
public static String getAllOcr(String file) throws Exception {
BufferedImage img = removeBackgroud(file);//去除重影
List<BufferedImage> listImg = splitImage(img);//切割圖片
Map<BufferedImage, String> map = loadTrainData();
String result = "";
for (BufferedImage bi : listImg) {
result += getSingleCharOcr(bi, map);
}
//ImageIO.write(img, "JPG", new File("result6\\" + result + ".jpg"));
return result;
}
/***
* 下載驗證碼圖片暫時保存供識別程序使用
* @param imgurl 驗證碼圖片url
*/
public static void downloadimg(String imgurl)
{
//HttpClient httpClient = new HttpClient();
//httpClient.getHttpConnectionManager().getParams().setConnectionTimeout(10000);
//httpClient.getHttpConnectionManager().getParams().setSoTimeout(10000);
GetMethod getMethod = new GetMethod(imgurl);
try {
int statusCode = client.executeMethod(getMethod);
System.out.println(statusCode);
if (statusCode != HttpStatus.SC_OK) {
System.err.println("("+statusCode+")Method failed: "+ getMethod.getStatusLine());
logger.info("("+statusCode+")Method failed: "+ getMethod.getStatusLine());
}
InputStream inputStream = getMethod.getResponseBodyAsStream();
OutputStream outStream = new FileOutputStream("/data/sata/share_sata/AmazonCrawl/amazonWeb/captcha.jpg");
IOUtils.copy(inputStream, outStream);
inputStream.close();
outStream.close();
} catch (IOException e) {
// TODO Auto-generated catch block
//logger.info(new Date()+"captcha appear exception:"+e.getMessage());
try {
//若遇到異常則睡眠20秒后繼續重試
Thread.sleep(20000);
} catch (InterruptedException e1) {
logger.error(e1);
}
e.printStackTrace();
}finally {
getMethod.releaseConnection();
}
}
/***
* 抽取頁面驗證碼並返回
* @param stringBuffer
* @return 驗證碼字符串
*/
public static String GetCaptcha(StringBuilder html){
String captcha_str="######";//未識別則為#
Document doc = Jsoup.parse(html.toString());
String imgurl = doc.select("div[class=a-row a-text-center]").get(0).child(0).attr("src");
//System.out.println(imgurl);
downloadimg(imgurl);
try {
captcha_str = getAllOcr("/data/sata/share_sata/AmazonCrawl/amazonWeb/captcha.jpg");
} catch (Exception e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
return captcha_str;
}
}
后記：復雜驗證碼識別
對於復雜的驗證碼識別：目前的最簡單的方案就是交給第三方人工打碼平台：可以參考我做的EBay多線程打碼兔驗證碼解決方案：
http://www.cnblogs.com/zeze/p/6402963.html
更專業的可以采用機器學習、模式識別等方法去實現，但是識別成功率，我目前測試的結果不是很理想，復雜的驗證碼，正確率在百分之二三十上下，但是我的訓練樣本庫不是很大，提高訓練的樣本可能結果會好一點。