public void PerformOcr(List<Tuple<int, int>> textRowLocations)
{
int DarkPixels;
var engine = new TesseractEngine(@"./tessdata", Form1.RegulatedNoiseSettings.TraineddataFile, EngineMode.Default);
engine.DefaultPageSegMode = PageSegMode.SingleLine;
string Stationname_OCR;
string StationameAnalysisBase;
// delete the old brainerous images - otherwise Brainerous will process older but not relevant images too
if (Directory.Exists(@".\Brainerous\images"))
foreach (string file in Directory.GetFiles(@".\\Brainerous\\images", "*.*"))
File.Delete(file);
else
Directory.CreateDirectory("./Brainerous/images");
float level;
var text = AnalyseFrameUsingTesseract(_bTrimmedHeader, engine, out level);
Stationname_OCR = StripPunctuationFromScannedText(text);// (text + " {" + page.GetMeanConfidence() + "}\r\n");
// debug: SystemAtTimeOfScreenshot = "BD+65 1846"
string[] StationsInSystem = _callingForm.myMilkyway.getStationNames(SystemAtTimeOfScreenshot);
if(_callingForm.tbCurrentStationinfoFromLogs.Text.Equals("scanning...", StringComparison.InvariantCultureIgnoreCase))
StationameAnalysisBase = Stationname_OCR;
else
StationameAnalysisBase = _callingForm.tbCurrentStationinfoFromLogs.Text;
string headerResult_temp = StationsInSystem.FirstOrDefault(x => x.Equals(StationameAnalysisBase, StringComparison.InvariantCultureIgnoreCase));
if(headerResult_temp == null)
{
// station not found in database
var matchesInStationReferenceList = StationsInSystem.OrderBy(x => _levenshtein.LD2(Stationname_OCR, x)).ToList();
if(matchesInStationReferenceList.Count > 0)
{
var ld = _levenshtein.LD2(Stationname_OCR, matchesInStationReferenceList[0].ToUpper());
// this depends on the length of the word - this factor works really good
double LevenshteinLimit = Math.Round((matchesInStationReferenceList[0].Length * 1.0), 0);
if (ld <= LevenshteinLimit)
Stationname_OCR = matchesInStationReferenceList[0];
}
}
else
{
Stationname_OCR = headerResult_temp;
}
// show station on GUI
_callingForm.DisplayResults(Stationname_OCR);
var commodityColumnText = new string[textRowLocations.Count(), 8];
var originalBitmaps = new Bitmap[textRowLocations.Count(),8];
var originalBitmapConfidences = new float[textRowLocations.Count(), 8];
var rowIds = new string[textRowLocations.Count()];
var rowCtr = 0;
var bitmapCtr = 0;
foreach (var row in textRowLocations)
{
int startRow = row.Item1 - 3;
int heightRow = row.Item2 - row.Item1 + 6;
if (startRow < 0)
startRow = 0;
if (heightRow + startRow > _bTrimmed_4_OCR.Height)
heightRow = _bTrimmed_4_OCR.Height - startRow;
// We'll use this later to identify the right correction image
rowIds[rowCtr] = Guid.NewGuid().ToString();
using (Bitmap b = RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(0, startRow, _bTrimmed_4_OCR.Width, heightRow)))
{
b.Save(".//OCR Correction Images//" + rowIds[rowCtr] + ".png");
}
int columnCounter = 0;
while (columnCounter < 8)
{
int left, width;
switch(columnCounter)
{
case 0:
// commodity
left = 0; width = _calibrationPoints[3].X - _calibrationPoints[2].X;
break;
case 1:
// sell
left = _calibrationPoints[3].X - _calibrationPoints[2].X; width = _calibrationPoints[4].X - _calibrationPoints[3].X;
break;
case 2:
//buy
left = _calibrationPoints[4].X - _calibrationPoints[2].X; width = _calibrationPoints[5].X - _calibrationPoints[4].X;
break;
case 3:
// freight
left = _calibrationPoints[5].X - _calibrationPoints[2].X; width = _calibrationPoints[6].X - _calibrationPoints[5].X;
break;
case 4:
// demand
left = _calibrationPoints[6].X - _calibrationPoints[2].X; width = _calibrationPoints[7].X - _calibrationPoints[6].X;
break;
case 5:
// demand level
left = _calibrationPoints[7].X - _calibrationPoints[2].X; width = _calibrationPoints[8].X - _calibrationPoints[7].X;
break;
case 6:
// supply
left = _calibrationPoints[8].X - _calibrationPoints[2].X; width = _calibrationPoints[9].X - _calibrationPoints[8].X;
break;
case 7:
// supply level
left = _calibrationPoints[9].X - _calibrationPoints[2].X; width = _calibrationPoints[10].X - _calibrationPoints[9].X;
break;
default:
left = 0; width = _calibrationPoints[3].X - _calibrationPoints[2].X;
break;
}
var fudgeFactor = 0;// _bOriginal.Height * 6 / 1440;
left = left + fudgeFactor;
width = width - fudgeFactor;
DarkPixels = 0;
if (_callingForm.cbCheckAOne.Checked)
{
if (PixelTest == null)
PixelTest = new EBPixeltest();
if (columnCounter == 3)
{
var brainerousOut = RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow, width, heightRow));
// check how much dark pixels are on the bitmap
for (int i = 0; i < brainerousOut.Height; i++)
for (int j = 0; j < brainerousOut.Width; j++)
if (brainerousOut.GetPixel(j, i).GetBrightness() < Form1.RegulatedNoiseSettings.EBPixelThreshold)
DarkPixels++;
PixelTest.addPicture(brainerousOut, DarkPixels);
}
}
else
{
// RNGraphics.Crop a little bit more form the left border because sometimes if theres
// the line of the table it was recognized as "1" or "7"
left += 10;
width -= 10;
if (columnCounter != 0 && columnCounter != 5 && columnCounter != 7)
{ //If it's a numeric column write it out for Brainerous to process later
var brainerousOut = RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow, width, heightRow));
if (Form1.RegulatedNoiseSettings.EBPixelAmount > 0)
{
// check how much dark pixels are on the bitmap -> we process only bitmaps
// with something on it (minimum one digit supposed, a "1" hat about 25 pixels in default 1920x1200)
for (int i = 0; i < brainerousOut.Height; i++)
for (int j = 0; j < brainerousOut.Width; j++)
if (brainerousOut.GetPixel(j, i).GetBrightness() < Form1.RegulatedNoiseSettings.EBPixelThreshold)
DarkPixels++;
}
if (DarkPixels >= Form1.RegulatedNoiseSettings.EBPixelAmount)
brainerousOut.Save("./Brainerous/images/" + bitmapCtr + ".png");
bitmapCtr++;
}
else
{ // It's a text column, we'll use Tesseract
// Prepare some different versions of the bitmap, we will take the best result
var c = new Bitmap[7];
c[0] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow, width, heightRow)));
c[1] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left + 1, startRow, width, heightRow)));
c[2] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left - 1, startRow, width, heightRow)));
c[3] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow - 1, width, heightRow)));
c[4] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left + 1, startRow - 1, width, heightRow)));
c[5] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left - 1, startRow - 1, width, heightRow)));
c[6] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow + 2, width, heightRow - 2)));
var t = new string[c.Length];
var cf = new float[c.Length];
for (int i = 0; i < c.Length; i++)
{
t[i] = AnalyseFrameUsingTesseract((Bitmap)(c[i].Clone()), engine, out cf[i]);
}
int result = 0;
float confidence = cf[0];
for (int i = 1; i < c.Length; i++)
{
if (confidence < cf[i])
{ result = i; confidence = cf[i]; }
}
originalBitmaps[rowCtr, columnCounter] = (Bitmap)(c[result].Clone());
switch (columnCounter)
{
//bodges for number columns
case 1:
case 2:
case 3:
t[result] = t[result].Replace(" ", "").Replace("O", "0").Replace("I", "1").Replace("'", "");
t[result] = System.Text.RegularExpressions.Regex.Replace(t[result], @"[a-zA-Z\s]+", string.Empty); // remove any alphas that remain
break;
case 5:
case 7:
var darkpixels = 0;
for (int i = 0; i < c[result].Height; i++)
for (int j = 0; j < c[result].Width; j++)
if (c[result].GetPixel(j, i).GetBrightness() < Form1.RegulatedNoiseSettings.EBPixelThreshold)
darkpixels++;
var ratio = (float) c[result].Height*c[result].Width/darkpixels;
if (darkpixels < 100)
{
//0
}
else if (ratio < 2)
{
//3
t[result] = "HIGH";
cf[result] = 1;
}
else if (ratio < 2.6)
{
//2
t[result] = "MED";
cf[result] = 1;
}
else
{
//1
t[result] = "LOW";
cf[result] = 1;
}
break;
}
if ((columnCounter == 5 && t[result].Contains("ENTER")) ||
(columnCounter == 6 && (t[result].Contains("NGAR") || t[result].Contains("SURFACE"))))
{
t[result] = "";
cf[result] = 1;
}
commodityColumnText[rowCtr, columnCounter] += t[result];
originalBitmapConfidences[rowCtr, columnCounter] = cf[result];
}
}
columnCounter++;
}
rowCtr++;
}
if (_callingForm.cbCheckAOne.Checked)
{
PixelTest.StartModal(_callingForm);
}
else
{
if (textRowLocations.Count > 0)
{
// Call out to Brainerous to process the numeric bitmaps we saved earlier
var outputFromBrainerous = "";
var pr = new Process();
pr.StartInfo.UseShellExecute = false;
pr.StartInfo.CreateNoWindow = true;
pr.StartInfo.RedirectStandardOutput = true;
pr.StartInfo.FileName = "./Brainerous/nn_training.exe";
pr.StartInfo.WorkingDirectory = "./Brainerous/";
pr.Start();
outputFromBrainerous = pr.StandardOutput.ReadToEnd();
while (outputFromBrainerous.Contains("Failed to pad successfully"))
{
var o2 = outputFromBrainerous.IndexOf("Failed to ");
var o3 = outputFromBrainerous.Substring(0, o2);
var o4 = outputFromBrainerous.Substring(o2).IndexOf("./images", StringComparison.InvariantCultureIgnoreCase);
// I had a string with "Failed to pad successfully" and only some trash behind but no "./images"
// so "o4" was "-1" and this results in strange behaviour
if (o4 > 0)
{
var o5 = outputFromBrainerous.Substring(o2 + o4);
outputFromBrainerous = o3 + "\r\n" + o5;
}
else
{
outputFromBrainerous = o3;
}
}
pr.WaitForExit();
List<string> splitOutput = ((string[])outputFromBrainerous.Replace("\r", "").Split('\n')).ToList();
for (var i = 0; i < (textRowLocations.Count * 10); i += 2)
{
string Filename = (i / 2).ToString() + ".png";
if ((splitOutput.Count <= i) || (splitOutput[i].Length < 14) || (splitOutput[i].Substring(9) != Filename))
{
splitOutput.Insert(i, "./images/" + Filename);
splitOutput.Insert(i + 1, "");
}
}
// Load the result from Brainerous into the OCR output
for (var i = 0; i < textRowLocations.Count; i++)
{
commodityColumnText[i, 1] = splitOutput[i * 10 + 1];
originalBitmaps[i, 1] = null;
originalBitmapConfidences[i, 1] = 1;
commodityColumnText[i, 2] = splitOutput[i * 10 + 3];
originalBitmaps[i, 2] = null;
originalBitmapConfidences[i, 2] = 1;
commodityColumnText[i, 3] = splitOutput[i * 10 + 5];
originalBitmaps[i, 3] = null;
originalBitmapConfidences[i, 3] = 1;
commodityColumnText[i, 4] = splitOutput[i * 10 + 7];
originalBitmaps[i, 4] = null;
originalBitmapConfidences[i, 4] = 1;
commodityColumnText[i, 6] = splitOutput[i * 10 + 9];
originalBitmaps[i, 6] = null;
originalBitmapConfidences[i, 6] = 1;
}
}
}
_bOriginal.Dispose();
_bOriginalClone.Dispose();
engine.Dispose();
if (_callingForm.cbCheckAOne.Checked)
{
_callingForm.setCheckbox(_callingForm.cbCheckAOne, false);
Form1.InstanceObject.clearOcrOutput();
}
else
{
// Send the results for this screenshot back to the Form
_callingForm.DisplayCommodityResults(commodityColumnText, originalBitmaps, originalBitmapConfidences, rowIds, CurrentScreenshot);
}
// ...and if we've got any buffered screenshots waiting to be processed, process the next one
if (ScreenshotBuffer.Count > 0)
{
var screenshot = ScreenshotBuffer[0];
ScreenshotBuffer.Remove(screenshot);
ProcessNewScreenshot(screenshot);
}
Working = false;
Debug.WriteLine("set to " + Working);
}
public void PerformOcr(List <Tuple <int, int> > textRowLocations)
{
int DarkPixels;
var engine = new TesseractEngine(@"./tessdata", Form1.RegulatedNoiseSettings.TraineddataFile, EngineMode.Default);
engine.DefaultPageSegMode = PageSegMode.SingleLine;
string Stationname_OCR;
string StationameAnalysisBase;
// delete the old brainerous images - otherwise Brainerous will process older but not relevant images too
if (Directory.Exists(@".\Brainerous\images"))
{
foreach (string file in Directory.GetFiles(@".\\Brainerous\\images", "*.*"))
{
File.Delete(file);
}
}
else
{
Directory.CreateDirectory("./Brainerous/images");
}
float level;
var text = AnalyseFrameUsingTesseract(_bTrimmedHeader, engine, out level);
Stationname_OCR = StripPunctuationFromScannedText(text);// (text + " {" + page.GetMeanConfidence() + "}\r\n");
// debug: SystemAtTimeOfScreenshot = "BD+65 1846"
string[] StationsInSystem = _callingForm.myMilkyway.getStationNames(SystemAtTimeOfScreenshot);
if (_callingForm.tbCurrentStationinfoFromLogs.Text.Equals("scanning...", StringComparison.InvariantCultureIgnoreCase))
{
StationameAnalysisBase = Stationname_OCR;
}
else
{
StationameAnalysisBase = _callingForm.tbCurrentStationinfoFromLogs.Text;
}
string headerResult_temp = StationsInSystem.FirstOrDefault(x => x.Equals(StationameAnalysisBase, StringComparison.InvariantCultureIgnoreCase));
if (headerResult_temp == null)
{
// station not found in database
var matchesInStationReferenceList = StationsInSystem.OrderBy(x => _levenshtein.LD2(Stationname_OCR, x)).ToList();
if (matchesInStationReferenceList.Count > 0)
{
var ld = _levenshtein.LD2(Stationname_OCR, matchesInStationReferenceList[0].ToUpper());
// this depends on the length of the word - this factor works really good
double LevenshteinLimit = Math.Round((matchesInStationReferenceList[0].Length * 1.0), 0);
if (ld <= LevenshteinLimit)
{
Stationname_OCR = matchesInStationReferenceList[0];
}
}
}
else
{
Stationname_OCR = headerResult_temp;
}
// show station on GUI
_callingForm.DisplayResults(Stationname_OCR);
var commodityColumnText = new string[textRowLocations.Count(), 8];
var originalBitmaps = new Bitmap[textRowLocations.Count(), 8];
var originalBitmapConfidences = new float[textRowLocations.Count(), 8];
var rowIds = new string[textRowLocations.Count()];
var rowCtr = 0;
var bitmapCtr = 0;
foreach (var row in textRowLocations)
{
int startRow = row.Item1 - 3;
int heightRow = row.Item2 - row.Item1 + 6;
if (startRow < 0)
{
startRow = 0;
}
if (heightRow + startRow > _bTrimmed_4_OCR.Height)
{
heightRow = _bTrimmed_4_OCR.Height - startRow;
}
// We'll use this later to identify the right correction image
rowIds[rowCtr] = Guid.NewGuid().ToString();
using (Bitmap b = RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(0, startRow, _bTrimmed_4_OCR.Width, heightRow)))
{
b.Save(".//OCR Correction Images//" + rowIds[rowCtr] + ".png");
}
int columnCounter = 0;
while (columnCounter < 8)
{
int left, width;
switch (columnCounter)
{
case 0:
// commodity
left = 0; width = _calibrationPoints[3].X - _calibrationPoints[2].X;
break;
case 1:
// sell
left = _calibrationPoints[3].X - _calibrationPoints[2].X; width = _calibrationPoints[4].X - _calibrationPoints[3].X;
break;
case 2:
//buy
left = _calibrationPoints[4].X - _calibrationPoints[2].X; width = _calibrationPoints[5].X - _calibrationPoints[4].X;
break;
case 3:
// freight
left = _calibrationPoints[5].X - _calibrationPoints[2].X; width = _calibrationPoints[6].X - _calibrationPoints[5].X;
break;
case 4:
// demand
left = _calibrationPoints[6].X - _calibrationPoints[2].X; width = _calibrationPoints[7].X - _calibrationPoints[6].X;
break;
case 5:
// demand level
left = _calibrationPoints[7].X - _calibrationPoints[2].X; width = _calibrationPoints[8].X - _calibrationPoints[7].X;
break;
case 6:
// supply
left = _calibrationPoints[8].X - _calibrationPoints[2].X; width = _calibrationPoints[9].X - _calibrationPoints[8].X;
break;
case 7:
// supply level
left = _calibrationPoints[9].X - _calibrationPoints[2].X; width = _calibrationPoints[10].X - _calibrationPoints[9].X;
break;
default:
left = 0; width = _calibrationPoints[3].X - _calibrationPoints[2].X;
break;
}
var fudgeFactor = 0;// _bOriginal.Height * 6 / 1440;
left = left + fudgeFactor;
width = width - fudgeFactor;
DarkPixels = 0;
if (_callingForm.cbCheckAOne.Checked)
{
if (PixelTest == null)
{
PixelTest = new EBPixeltest();
}
if (columnCounter == 3)
{
var brainerousOut = RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow, width, heightRow));
// check how much dark pixels are on the bitmap
for (int i = 0; i < brainerousOut.Height; i++)
{
for (int j = 0; j < brainerousOut.Width; j++)
{
if (brainerousOut.GetPixel(j, i).GetBrightness() < Form1.RegulatedNoiseSettings.EBPixelThreshold)
{
DarkPixels++;
}
}
}
PixelTest.addPicture(brainerousOut, DarkPixels);
}
}
else
{
// RNGraphics.Crop a little bit more form the left border because sometimes if theres
// the line of the table it was recognized as "1" or "7"
left += 10;
width -= 10;
if (columnCounter != 0 && columnCounter != 5 && columnCounter != 7)
{ //If it's a numeric column write it out for Brainerous to process later
var brainerousOut = RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow, width, heightRow));
if (Form1.RegulatedNoiseSettings.EBPixelAmount > 0)
{
// check how much dark pixels are on the bitmap -> we process only bitmaps
// with something on it (minimum one digit supposed, a "1" hat about 25 pixels in default 1920x1200)
for (int i = 0; i < brainerousOut.Height; i++)
{
for (int j = 0; j < brainerousOut.Width; j++)
{
if (brainerousOut.GetPixel(j, i).GetBrightness() < Form1.RegulatedNoiseSettings.EBPixelThreshold)
{
DarkPixels++;
}
}
}
}
if (DarkPixels >= Form1.RegulatedNoiseSettings.EBPixelAmount)
{
brainerousOut.Save("./Brainerous/images/" + bitmapCtr + ".png");
}
bitmapCtr++;
}
else
{ // It's a text column, we'll use Tesseract
// Prepare some different versions of the bitmap, we will take the best result
var c = new Bitmap[7];
c[0] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow, width, heightRow)));
c[1] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left + 1, startRow, width, heightRow)));
c[2] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left - 1, startRow, width, heightRow)));
c[3] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow - 1, width, heightRow)));
c[4] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left + 1, startRow - 1, width, heightRow)));
c[5] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left - 1, startRow - 1, width, heightRow)));
c[6] = (RNGraphics.Crop(_bTrimmed_4_OCR, new Rectangle(left, startRow + 2, width, heightRow - 2)));
var t = new string[c.Length];
var cf = new float[c.Length];
for (int i = 0; i < c.Length; i++)
{
t[i] = AnalyseFrameUsingTesseract((Bitmap)(c[i].Clone()), engine, out cf[i]);
}
int result = 0;
float confidence = cf[0];
for (int i = 1; i < c.Length; i++)
{
if (confidence < cf[i])
{
result = i; confidence = cf[i];
}
}
originalBitmaps[rowCtr, columnCounter] = (Bitmap)(c[result].Clone());
switch (columnCounter)
{
//bodges for number columns
case 1:
case 2:
case 3:
t[result] = t[result].Replace(" ", "").Replace("O", "0").Replace("I", "1").Replace("'", "");
t[result] = System.Text.RegularExpressions.Regex.Replace(t[result], @"[a-zA-Z\s]+", string.Empty); // remove any alphas that remain
break;
case 5:
case 7:
t[result] = t[result].Replace(" ", "").Replace("-", "");
if (t[result] == "HIGH" || t[result] == "MED" || t[result] == "LOW")
{
cf[result] = 1;
}
break;
}
if ((columnCounter == 5 && t[result].Contains("ENTER")) ||
(columnCounter == 6 && (t[result].Contains("NGAR") || t[result].Contains("SURFACE"))))
{
t[result] = "";
cf[result] = 1;
}
commodityColumnText[rowCtr, columnCounter] += t[result];
originalBitmapConfidences[rowCtr, columnCounter] = cf[result];
}
}
columnCounter++;
}
rowCtr++;
}
if (_callingForm.cbCheckAOne.Checked)
{
PixelTest.StartModal(_callingForm);
}
else
{
if (textRowLocations.Count > 0)
{
// Call out to Brainerous to process the numeric bitmaps we saved earlier
var outputFromBrainerous = "";
var pr = new Process();
pr.StartInfo.UseShellExecute = false;
pr.StartInfo.CreateNoWindow = true;
pr.StartInfo.RedirectStandardOutput = true;
pr.StartInfo.FileName = "./Brainerous/nn_training.exe";
pr.StartInfo.WorkingDirectory = "./Brainerous/";
pr.Start();
outputFromBrainerous = pr.StandardOutput.ReadToEnd();
while (outputFromBrainerous.Contains("Failed to pad successfully"))
{
var o2 = outputFromBrainerous.IndexOf("Failed to ");
var o3 = outputFromBrainerous.Substring(0, o2);
var o4 = outputFromBrainerous.Substring(o2).IndexOf("./images", StringComparison.InvariantCultureIgnoreCase);
// I had a string with "Failed to pad successfully" and only some trash behind but no "./images"
// so "o4" was "-1" and this results in strange behaviour
if (o4 > 0)
{
var o5 = outputFromBrainerous.Substring(o2 + o4);
outputFromBrainerous = o3 + "\r\n" + o5;
}
else
{
outputFromBrainerous = o3;
}
}
pr.WaitForExit();
List <string> splitOutput = ((string[])outputFromBrainerous.Replace("\r", "").Split('\n')).ToList();
for (var i = 0; i < (textRowLocations.Count * 10); i += 2)
{
string Filename = (i / 2).ToString() + ".png";
if ((splitOutput.Count <= i) || (splitOutput[i].Length < 14) || (splitOutput[i].Substring(9) != Filename))
{
splitOutput.Insert(i, "./images/" + Filename);
splitOutput.Insert(i + 1, "");
}
}
// Load the result from Brainerous into the OCR output
for (var i = 0; i < textRowLocations.Count; i++)
{
commodityColumnText[i, 1] = splitOutput[i * 10 + 1];
originalBitmaps[i, 1] = null;
originalBitmapConfidences[i, 1] = 1;
commodityColumnText[i, 2] = splitOutput[i * 10 + 3];
originalBitmaps[i, 2] = null;
originalBitmapConfidences[i, 2] = 1;
commodityColumnText[i, 3] = splitOutput[i * 10 + 5];
originalBitmaps[i, 3] = null;
originalBitmapConfidences[i, 3] = 1;
commodityColumnText[i, 4] = splitOutput[i * 10 + 7];
originalBitmaps[i, 4] = null;
originalBitmapConfidences[i, 4] = 1;
commodityColumnText[i, 6] = splitOutput[i * 10 + 9];
originalBitmaps[i, 6] = null;
originalBitmapConfidences[i, 6] = 1;
}
}
}
_bOriginal.Dispose();
_bOriginalClone.Dispose();
engine.Dispose();
if (_callingForm.cbCheckAOne.Checked)
{
_callingForm.setCheckbox(_callingForm.cbCheckAOne, false);
Form1.InstanceObject.clearOcrOutput();
}
else
{
// Send the results for this screenshot back to the Form
_callingForm.DisplayCommodityResults(commodityColumnText, originalBitmaps, originalBitmapConfidences, rowIds, CurrentScreenshot);
}
// ...and if we've got any buffered screenshots waiting to be processed, process the next one
if (ScreenshotBuffer.Count > 0)
{
var screenshot = ScreenshotBuffer[0];
ScreenshotBuffer.Remove(screenshot);
ProcessNewScreenshot(screenshot);
}
Working = false;
Debug.WriteLine("set to " + Working);
}