static void __executeBackground(byte[] buf)
{
oTesseractRequest r = null;
string guid = Encoding.ASCII.GetString(buf);
var redis = new RedisBase(new RedisSetting(REDIS_TYPE.ONLY_READ, 1000));
try
{
string json = redis.HGET("_OCR_REQUEST", guid);
r = JsonConvert.DeserializeObject <oTesseractRequest>(json);
Bitmap bitmap = redis.HGET_BITMAP(r.redis_key, r.redis_field);
if (bitmap != null)
{
r = __ocrExecute(r, bitmap, redis);
}
}
catch (Exception ex)
{
if (r != null)
{
string error = ex.Message + Environment.NewLine + ex.StackTrace
+ Environment.NewLine + "----------------" + Environment.NewLine +
JsonConvert.SerializeObject(r);
r.ok = -1;
redis.HSET("_OCR_REQ_ERR", r.requestId, error);
}
}
if (r != null)
{
redis.HSET("_OCR_REQUEST", r.requestId, JsonConvert.SerializeObject(r, Formatting.Indented));
redis.HSET("_OCR_REQ_LOG", r.requestId, r.ok.ToString());
redis.PUBLISH("__TESSERACT_OUT", r.requestId);
}
}
static oTesseractRequest __ocrExecute2(oTesseractRequest req, Bitmap bitmap)
{
using (TesseractProcessor processor = new TesseractProcessor())
{
processor.InitForAnalysePage();
using (GreyImage greyImage = GreyImage.FromImage(bitmap))
{
//greyImage.Save(ImageFormat.Bmp, outFile2);
ImageThresholder thresholder = new AdaptiveThresholder();
using (BinaryImage binImage = thresholder.Threshold(greyImage))
{
DocumentLayout doc = null;
switch (req.command)
{
case TESSERACT_COMMAND.GET_TEXT:
//string s = tes.GetText().Trim();
//req.output_text = s;
//req.output_count = s.Length;
req.ok = 1;
break;
default:
unsafe
{
doc = processor.AnalyseLayoutBinaryImage(
binImage.BinaryData, greyImage.Width, greyImage.Height);
}
if (doc != null)
{
var bs = new List <string>();
if (doc.Blocks.Count > 0)
{
for (int i = 0; i < doc.Blocks.Count; i++)
{
for (int j = 0; j < doc.Blocks[i].Paragraphs.Count; j++)
{
bs.AddRange(doc.Blocks[j].Paragraphs[j].Lines
.Select(x => string.Format(
"{0}_{1}_{2}_{3}", x.Left, x.Top, x.Right, x.Bottom)));
}
}
}
req.output_format = "left_top_right_bottom";
req.output_text = string.Join("|", bs.ToArray());
req.output_count = bs.Count;
req.ok = 1;
}
break;
}
}
}
}
return(req);
}
static oTesseractRequest __ocrExecute(oTesseractRequest req, Bitmap image)
{
PageIteratorLevel level = PageIteratorLevel.Word;
switch (req.command)
{
case TESSERACT_COMMAND.GET_SEGMENTED_REGION_BLOCK:
level = PageIteratorLevel.Block;
break;
case TESSERACT_COMMAND.GET_SEGMENTED_REGION_PARA:
level = PageIteratorLevel.Para;
break;
case TESSERACT_COMMAND.GET_SEGMENTED_REGION_SYMBOL:
level = PageIteratorLevel.Symbol;
break;
case TESSERACT_COMMAND.GET_SEGMENTED_REGION_TEXTLINE:
level = PageIteratorLevel.TextLine;
break;
case TESSERACT_COMMAND.GET_SEGMENTED_REGION_WORD:
level = PageIteratorLevel.Word;
break;
case TESSERACT_COMMAND.GET_TEXT:
break;
}
EngineMode mode = EngineMode.Default;
switch (req.mode)
{
case ENGINE_MODE.LSTM_ONLY:
mode = EngineMode.LstmOnly;
break;
case ENGINE_MODE.TESSERACT_AND_LSTM:
mode = EngineMode.TesseractAndLstm;
break;
case ENGINE_MODE.TESSERACT_ONLY:
mode = EngineMode.TesseractOnly;
break;
}
using (var engine = new TesseractEngine(req.data_path, req.lang, mode))
using (var pix = new BitmapToPixConverter().Convert(image))
{
using (var tes = engine.Process(pix))
{
switch (req.command)
{
case TESSERACT_COMMAND.GET_TEXT:
string s = tes.GetText().Trim();
req.output_text = s;
req.output_count = s.Length;
req.ok = 1;
break;
default:
var boxes = tes.GetSegmentedRegions(level).Select(x =>
string.Format("{0}_{1}_{2}_{3}", x.X, x.Y, x.Width, x.Height)).ToArray();
req.output_format = "x_y_width_height";
req.output_text = string.Join("|", boxes.Select(x => x.ToString()).ToArray());
req.output_count = boxes.Length;
req.ok = 1;
break;
}
}
}
return(req);
}
static oTesseractRequest __ocrExecute(oTesseractRequest req, Bitmap bitmap)
{
using (TesseractProcessor processor = new TesseractProcessor())
{
// call SetVariable() method before passing image(api->SetImage(image))
// 0: otsu
// 1: isodata local adaptive
// 2: sauvola local adaptive => not implement yet
//processor.SetVariable("tessedit_thresholding_method", "0");
processor.SetVariable("tessedit_thresholding_method", "1");
//processor.InitForAnalysePage();
//processor.SetPageSegMode(ePageSegMode.PSM_AUTO_ONLY);
//var success = processor.Init(req.data_path, req.lang, (int)eOcrEngineMode.OEM_DEFAULT);
//var imageColor = new Emgu.CV.Mat();
//var imageCV = new Emgu.CV.Image<Emgu.CV.Structure.Bgr, byte>(bitmap); //Image Class from Emgu.CV
////var imageCV = new Emgu.CV.Image<Emgu.CV.Structure.Gray, byte>(bitmap); //Image Class from Emgu.CV
//var image = imageCV.Mat; //This is your Image converted to Mat
//if (image.NumberOfChannels == 1)
// Emgu.CV.CvInvoke.CvtColor(image, imageColor, Emgu.CV.CvEnum.ColorConversion.Gray2Bgr);
//else
// image.CopyTo(imageColor);
using (var m0 = new MemoryStream())
{
bitmap.Save(m0, ImageFormat.Jpeg);
//using (Bitmap bmp = Bitmap.FromFile(@"C:\temp\1.jpg") as Bitmap)
using (Bitmap bmp = new Bitmap(m0))
{
DocumentLayout doc = null;
switch (req.command)
{
case TESSERACT_COMMAND.GET_TEXT:
//string s = tes.GetText().Trim();
//req.output_text = s;
//req.output_count = s.Length;
req.ok = 1;
break;
default:
unsafe
{
doc = processor.AnalyseLayout(bmp);
}
if (doc != null)
{
var bs = new List <string>();
if (doc.Blocks.Count > 0)
{
for (int i = 0; i < doc.Blocks.Count; i++)
{
for (int j = 0; j < doc.Blocks[i].Paragraphs.Count; j++)
{
bs.AddRange(doc.Blocks[j].Paragraphs[j].Lines
.Select(x => string.Format(
"{0}_{1}_{2}_{3}", x.Left, x.Top, x.Right, x.Bottom)));
}
}
}
req.output_format = "left_top_right_bottom";
req.output_text = string.Join("|", bs.ToArray());
req.output_count = bs.Count;
req.ok = 1;
}
break;
}
}
}
}
return(req);
}
