detailed
Download Code: http://www.demodashi.com/demo/11963.html
A few days ago in the afternoon nothing, there is a demand for a friend, said to identify the identity card number above, just idle, to help him solve it, do not say more than perfect, but at least be considered to address the needs, well, gossip less.
My first look at DEMO bar
Next we introduce one by one
First, identify networking
Is cut out from someone else's Demo, in fact, an online interface is also used someone else's, but I looked at should be considered "non-normal call" (this means we all understand it). Following analysis of the advantages and disadvantages of this method it.
Pros: fast, upload photos, will return all the information on the ID card, including name and address of birth, etc.
Cons: "non-normal" call to have a certain unreliability, if someday people off or changed the interface is more awkward, of course, you can choose to buy the official version of the people.
Second, local recognition
Tess_two do recognize based on this we can be assured. Probably look at how to use it!
First reference:
|
1
|
compile
'com.rmtheis:tess-two:6.0.0'
|
Then, in fact, very simple to use, but points to note
1. To have an SD card in his recognition library that you can be understood as a dictionary that you can train yourself, because I was trained by someone else (only consist of English and numbers), so I will not say how training , and Baidu have a lot of it.
2. It should be noted that, let him dictionary path of the folder name must be "tessdata", otherwise an error
Well, ready to do a good job, then introduces how to use me directly attached to the core code, the code notes, read messages or private letter I
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
|
//训练数据路径，tessdata
static
final
String TESSBASE_PATH = Environment.getExternalStorageDirectory() +
"/"
;
//识别语言英文
static
final
String DEFAULT_LANGUAGE =
"eng"
;
/**
* 传SD卡图片路径（当然你们也可以传Bitmap）
* @param url
*/
private
void
localre(String url) {
//把图片转为Bitmap
Bitmap bmp = BitmapFactory.decodeFile(url);
//创建Tess
final
TessBaseAPI baseApi =
new
TessBaseAPI();
//下面这一块代码为裁取身份证号码区域（否则识别乱码，不准确）
int
x, y, w, h;
x = (
int
) (bmp.getWidth() *
0.340
);
y = (
int
) (bmp.getHeight() *
0.800
);
w = (
int
) (bmp.getWidth() *
0.6
+
0
.5f);
h = (
int
) (bmp.getHeight() *
0.12
+
0
.5f);
Bitmap bit_hm = Bitmap.createBitmap(bmp, x, y, w, h);
//这个只是我将裁取的号码区展示在了一个ImageView上，这个可以没有
iv_number.setImageBitmap(bit_hm);
//初始化OCR的训练数据路径与语言
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
//设置识别模式
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
//设置要识别的图片
baseApi.setImage(bit_hm);
//设置字典白名单
baseApi.setVariable(
"tessedit_char_whitelist"
,
"0123456789Xx"
);
//把识别内容设置到EditText里
tv_result.setText(baseApi.getUTF8Text());
//收尾
baseApi.clear();
baseApi.end();
}
|
OK,就这么简单，图片清晰切裁取区域正确的情况下，准确度几乎100%；
给大家举个身份证照片的例子吧，否则裁取号码会不
上一张结果图
实时识别
其实就是本地识别的拓展版，把摄像头的数据转为Bitmap,去识别，还是贴核心代码吧，看不懂的自己下Demo研究。
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
|
/**
* 摄像头数据回调
* @param data
* @param camera
*/
@Override
public
void
onPreviewFrame(
byte
[] data, Camera camera) {
camera.addCallbackBuffer(data);
//将byte数组转为Bitmap
ByteArrayOutputStream baos;
byte
[] rawImage;
Bitmap bitmap;
Camera.Size previewSize = camera.getParameters().getPreviewSize();
//获取尺寸,格式转换的时候要用到
BitmapFactory.Options newOpts =
new
BitmapFactory.Options();
newOpts.inJustDecodeBounds =
true
;
YuvImage yuvimage =
new
YuvImage(
data,
ImageFormat.NV21,
previewSize.width,
previewSize.height,
null
);
baos =
new
ByteArrayOutputStream();
yuvimage.compressToJpeg(
new
Rect(
0
,
0
, previewSize.width, previewSize.height),
100
, baos);
// 80--JPG图片的质量[0-100],100最高
rawImage = baos.toByteArray();
//将rawImage转换成bitmap
BitmapFactory.Options options =
new
BitmapFactory.Options();
options.inPreferredConfig = Bitmap.Config.RGB_565;
bitmap = BitmapFactory.decodeByteArray(rawImage,
0
, rawImage.length, options);
if
(bitmap ==
null
) {
Log.d(
"zka"
,
"bitmap is nlll"
);
return
;
}
else
{
//裁取图片中央身份证区域
int
height = bitmap.getHeight();
int
width = bitmap.getWidth();
final
Bitmap bitmap1 = Bitmap.createBitmap(bitmap, width/
2
- dip2px(
150
),height /
2
- dip2px(
92
), dip2px(
300
), dip2px(
185
));
//截取身份证号码区域
int
x, y, w, h;
x = (
int
) (bitmap1.getWidth() *
0.340
);
y = (
int
) (bitmap1.getHeight() *
0.800
);
w = (
int
) (bitmap1.getWidth() *
0.6
+
0
.5f);
h = (
int
) (bitmap1.getHeight() *
0.12
+
0
.5f);
Bitmap bit_hm = Bitmap.createBitmap(bitmap1, x, y, w, h);
// 识别
if
(bit_hm !=
null
){
String localre = localre(bit_hm);
if
(localre.length() ==
18
) {
Log.e(TAG,
"onPreviewFrame: "
+localre );
Toast.makeText(getApplicationContext(),localre,Toast.LENGTH_SHORT).show();
}
}
}
}
/**
* 识别
* @param bm
* @return
*/
private
String localre(Bitmap bm) {
String content =
""
;
bm = bm.copy(Bitmap.Config.ARGB_8888,
true
);
iv_result.setImageBitmap(bm);
TessBaseAPI baseApi =
new
TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
//设置识别模式
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
//设置要识别的图片
baseApi.setImage(bm);
baseApi.setVariable(
"tessedit_char_whitelist"
,
"0123456789Xx"
);
Log.e(TAG,
"localre: "
+ baseApi.getUTF8Text());
content = baseApi.getUTF8Text();
baseApi.clear();
baseApi.end();
return
content;
}
|
三、源码包截图
四、其他
Ok，就这样吧！核心也就这些东西，有问题的可以留言或私信，有好的解决办法也可以交流，，出于隐私，就把人家的信息打码， 不过识别出来准确度是100%。