summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorNick White <git@njw.name>2019-05-14 10:29:29 +0100
committerNick White <git@njw.name>2019-05-14 10:29:29 +0100
commit3b12c59f37f0dc58ad1886052c03f4c81a2a5b78 (patch)
tree284c4a5edc03f932aa096ba84936f210122e1bb1
parent15b05839f91b29319b9d6fd475007ae3ad6224e2 (diff)
Add preprocmulti tool, that outputs multiple binarisation options quickly
-rw-r--r--preproc/cmd/preproc/main.go4
-rw-r--r--preproc/cmd/preprocmulti/main.go96
-rw-r--r--preproc/sauvola.go21
3 files changed, 119 insertions, 2 deletions
diff --git a/preproc/cmd/preproc/main.go b/preproc/cmd/preproc/main.go
index 123895f..92e8509 100644
--- a/preproc/cmd/preproc/main.go
+++ b/preproc/cmd/preproc/main.go
@@ -23,7 +23,7 @@ func autowsize(bounds image.Rectangle) int {
func main() {
flag.Usage = func() {
- fmt.Fprintf(os.Stderr, "Usage: preproc [-bt bintype] [-bw winsize] [-k num] [-t thresh] [-ws wipesize] inimg outimg\n")
+ fmt.Fprintf(os.Stderr, "Usage: preproc [-bt bintype] [-bw winsize] [-k num] [-wt wipethresh] [-ws wipesize] inimg outimg\n")
fmt.Fprintf(os.Stderr, "Binarize and preprocess an image\n")
flag.PrintDefaults()
}
@@ -31,7 +31,7 @@ func main() {
ksize := flag.Float64("k", 0.5, "K for sauvola binarization algorithm. This controls the overall threshold level. Set it lower for very light text (try 0.1 or 0.2).")
btype := flag.String("bt", "binary", "Type of binarization threshold. binary or zeroinv are currently implemented.")
wipewsize := flag.Int("ws", 5, "Window size for wiping algorithm.")
- thresh := flag.Float64("t", 0.05, "Threshold for the proportion of black pixels below which a window is determined to be the edge.")
+ thresh := flag.Float64("wt", 0.05, "Threshold for the wiping algorithm to determine the proportion of black pixels below which a window is determined to be the edge.")
flag.Parse()
if flag.NArg() < 2 {
flag.Usage()
diff --git a/preproc/cmd/preprocmulti/main.go b/preproc/cmd/preprocmulti/main.go
new file mode 100644
index 0000000..a159938
--- /dev/null
+++ b/preproc/cmd/preprocmulti/main.go
@@ -0,0 +1,96 @@
+package main
+
+// TODO: come up with a way to set a good ksize automatically
+// TODO: add minimum size variable (default ~30%?) for wipe
+
+import (
+ "flag"
+ "fmt"
+ "image"
+ "image/draw"
+ _ "image/jpeg"
+ "image/png"
+ "log"
+ "os"
+
+ "rescribe.xyz/go.git/integralimg"
+ "rescribe.xyz/go.git/preproc"
+)
+
+// TODO: do more testing to see how good this assumption is
+func autowsize(bounds image.Rectangle) int {
+ return bounds.Dx() / 60
+}
+
+func main() {
+ flag.Usage = func() {
+ fmt.Fprintf(os.Stderr, "Usage: preproc [-bt bintype] [-bw winsize] [-t thresh] [-ws wipesize] inimg outbase\n")
+ fmt.Fprintf(os.Stderr, "Binarize and preprocess an image, with multiple binarisation levels,\n")
+ fmt.Fprintf(os.Stderr, "saving images to outbase_knum.png.\n")
+ flag.PrintDefaults()
+ }
+ binwsize := flag.Int("bw", 0, "Window size for sauvola binarization algorithm. Set automatically based on resolution if not set.")
+ btype := flag.String("bt", "binary", "Type of binarization threshold. binary or zeroinv are currently implemented.")
+ wipewsize := flag.Int("ws", 5, "Window size for wiping algorithm.")
+ thresh := flag.Float64("t", 0.05, "Threshold for the proportion of black pixels below which a window is determined to be the edge.")
+ flag.Parse()
+ if flag.NArg() < 2 {
+ flag.Usage()
+ os.Exit(1)
+ }
+
+ log.Printf("Opening %s\n", flag.Arg(0))
+ f, err := os.Open(flag.Arg(0))
+ defer f.Close()
+ if err != nil {
+ log.Fatalf("Could not open file %s: %v\n", flag.Arg(0), err)
+ }
+ img, _, err := image.Decode(f)
+ if err != nil {
+ log.Fatalf("Could not decode image: %v\n", err)
+ }
+ b := img.Bounds()
+ gray := image.NewGray(image.Rect(0, 0, b.Dx(), b.Dy()))
+ draw.Draw(gray, b, img, b.Min, draw.Src)
+
+ if *binwsize == 0 {
+ *binwsize = autowsize(b)
+ }
+
+ if *binwsize % 2 == 0 {
+ *binwsize++
+ }
+
+ ksizes := []float64{0.2, 0.3, 0.4, 0.5, 0.6}
+
+ var threshimg image.Image
+ log.Print("Precalculating integral images")
+ integrals := integralimg.ToAllIntegralImg(gray)
+
+ for _, k := range ksizes {
+ log.Print("Binarising")
+ threshimg = preproc.PreCalcedSauvola(integrals, gray, k, *binwsize)
+
+ if *btype == "zeroinv" {
+ threshimg, err = preproc.BinToZeroInv(threshimg.(*image.Gray), img.(*image.RGBA))
+ if err != nil {
+ log.Fatal(err)
+ }
+ }
+
+ log.Print("Wiping sides")
+ clean := preproc.Wipe(threshimg.(*image.Gray), *wipewsize, *thresh)
+
+ savefn := fmt.Sprintf("%s_%0.1f.png", flag.Arg(1), k)
+ log.Printf("Saving %s\n", savefn)
+ f, err = os.Create(savefn)
+ if err != nil {
+ log.Fatalf("Could not create file %s: %v\n", savefn, err)
+ }
+ defer f.Close()
+ err = png.Encode(f, clean)
+ if err != nil {
+ log.Fatalf("Could not encode image: %v\n", err)
+ }
+ }
+}
diff --git a/preproc/sauvola.go b/preproc/sauvola.go
index e93ea81..2ed9a92 100644
--- a/preproc/sauvola.go
+++ b/preproc/sauvola.go
@@ -53,3 +53,24 @@ func IntegralSauvola(img *image.Gray, ksize float64, windowsize int) *image.Gray
return new
}
+
+// PreCalcedSauvola Implements Sauvola's algorithm using precalculated Integral Images
+// TODO: have this be the root function that the other two reference
+func PreCalcedSauvola(integrals integralimg.WithSq, img *image.Gray, ksize float64, windowsize int) *image.Gray {
+ b := img.Bounds()
+ new := image.NewGray(b)
+
+ for y := b.Min.Y; y < b.Max.Y; y++ {
+ for x := b.Min.X; x < b.Max.X; x++ {
+ m, dev := integrals.MeanStdDevWindow(x, y, windowsize)
+ threshold := m * (1 + ksize * ((dev / 128) - 1))
+ if img.GrayAt(x, y).Y < uint8(threshold) {
+ new.SetGray(x, y, color.Gray{0})
+ } else {
+ new.SetGray(x, y, color.Gray{255})
+ }
+ }
+ }
+
+ return new
+}