summaryrefslogtreecommitdiff
path: root/preproc/cmd
diff options
context:
space:
mode:
Diffstat (limited to 'preproc/cmd')
-rw-r--r--preproc/cmd/binarize/main.go78
-rw-r--r--preproc/cmd/preproc/main.go90
-rw-r--r--preproc/cmd/preprocmulti/main.go101
-rw-r--r--preproc/cmd/wipe/main.go55
4 files changed, 0 insertions, 324 deletions
diff --git a/preproc/cmd/binarize/main.go b/preproc/cmd/binarize/main.go
deleted file mode 100644
index e7f677e..0000000
--- a/preproc/cmd/binarize/main.go
+++ /dev/null
@@ -1,78 +0,0 @@
-package main
-
-import (
- "flag"
- "fmt"
- "image"
- "image/draw"
- _ "image/jpeg"
- "image/png"
- "log"
- "os"
-
- "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: binarize [-k num] [-t type] [-w num] inimg outimg\n")
- flag.PrintDefaults()
- }
- wsize := flag.Int("w", 0, "Window size for sauvola algorithm. Set automatically based on resolution if not set.")
- ksize := flag.Float64("k", 0.5, "K for sauvola algorithm. This controls the overall threshold level. Set it lower for very light text (try 0.1 or 0.2).")
- btype := flag.String("t", "binary", "Type of threshold. binary or zeroinv are currently implemented.")
- flag.Parse()
- if flag.NArg() < 2 {
- flag.Usage()
- os.Exit(1)
- }
-
- 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 *wsize == 0 {
- *wsize = autowsize(b)
- log.Printf("Set window size to %d\n", *wsize)
- }
-
- if *wsize%2 == 0 {
- *wsize++
- }
-
- // TODO: come up with a way to set a good ksize automatically
-
- var thresh image.Image
- thresh = preproc.IntegralSauvola(gray, *ksize, *wsize)
-
- if *btype == "zeroinv" {
- thresh, err = preproc.BinToZeroInv(thresh.(*image.Gray), img.(*image.RGBA))
- if err != nil {
- log.Fatal(err)
- }
- }
-
- f, err = os.Create(flag.Arg(1))
- if err != nil {
- log.Fatalf("Could not create file %s: %v\n", flag.Arg(1), err)
- }
- defer f.Close()
- err = png.Encode(f, thresh)
- if err != nil {
- log.Fatalf("Could not encode image: %v\n", err)
- }
-}
diff --git a/preproc/cmd/preproc/main.go b/preproc/cmd/preproc/main.go
deleted file mode 100644
index 1c248e0..0000000
--- a/preproc/cmd/preproc/main.go
+++ /dev/null
@@ -1,90 +0,0 @@
-package main
-
-// TODO: come up with a way to set a good ksize automatically
-
-import (
- "flag"
- "fmt"
- "image"
- "image/draw"
- _ "image/jpeg"
- "image/png"
- "log"
- "os"
-
- "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] [-k num] [-m minperc] [-nowipe] [-wt wipethresh] [-ws wipesize] inimg outimg\n")
- fmt.Fprintf(os.Stderr, "Binarize and preprocess an image\n")
- flag.PrintDefaults()
- }
- binwsize := flag.Int("bw", 0, "Window size for sauvola binarization algorithm. Set automatically based on resolution if not set.")
- 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.")
- min := flag.Int("m", 30, "Minimum percentage of the image width for the content width calculation to be considered valid.")
- nowipe := flag.Bool("nowipe", false, "Disable wiping completely.")
- wipewsize := flag.Int("ws", 5, "Window size for wiping algorithm.")
- 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()
- os.Exit(1)
- }
-
- 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++
- }
-
- log.Print("Binarising")
- var clean, threshimg image.Image
- threshimg = preproc.IntegralSauvola(gray, *ksize, *binwsize)
-
- if *btype == "zeroinv" {
- threshimg, err = preproc.BinToZeroInv(threshimg.(*image.Gray), img.(*image.RGBA))
- if err != nil {
- log.Fatal(err)
- }
- }
-
- if !*nowipe {
- log.Print("Wiping sides")
- clean = preproc.Wipe(threshimg.(*image.Gray), *wipewsize, *thresh, *min)
- } else {
- clean = threshimg
- }
-
- f, err = os.Create(flag.Arg(1))
- if err != nil {
- log.Fatalf("Could not create file %s: %v\n", flag.Arg(1), 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/cmd/preprocmulti/main.go b/preproc/cmd/preprocmulti/main.go
deleted file mode 100644
index c6c9fe4..0000000
--- a/preproc/cmd/preprocmulti/main.go
+++ /dev/null
@@ -1,101 +0,0 @@
-package main
-
-// TODO: come up with a way to set a good ksize automatically
-
-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() {
- ksizes := []float64{0.1, 0.2, 0.4, 0.5}
-
- flag.Usage = func() {
- fmt.Fprintf(os.Stderr, "Usage: preprocmulti [-bt bintype] [-bw winsize] [-m minperc] [-nowipe] [-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_bin{k}.png.\n")
- fmt.Fprintf(os.Stderr, "Binarises with these levels for k: %v.\n", ksizes)
- 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.")
- min := flag.Int("m", 30, "Minimum percentage of the image width for the content width calculation to be considered valid.")
- nowipe := flag.Bool("nowipe", false, "Disable wiping completely.")
- wipewsize := flag.Int("ws", 5, "Window size for wiping algorithm.")
- 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++
- }
-
- var clean, 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)
- }
- }
-
- if !*nowipe {
- log.Print("Wiping sides")
- clean = preproc.Wipe(threshimg.(*image.Gray), *wipewsize, k*0.02, *min)
- } else {
- clean = threshimg
- }
-
- savefn := fmt.Sprintf("%s_bin%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/cmd/wipe/main.go b/preproc/cmd/wipe/main.go
deleted file mode 100644
index e5c039d..0000000
--- a/preproc/cmd/wipe/main.go
+++ /dev/null
@@ -1,55 +0,0 @@
-package main
-
-import (
- "flag"
- "fmt"
- "image"
- "image/draw"
- _ "image/jpeg"
- "image/png"
- "log"
- "os"
-
- "rescribe.xyz/go.git/preproc"
-)
-
-func main() {
- flag.Usage = func() {
- fmt.Fprintf(os.Stderr, "Usage: wipe [-m minperc] [-t thresh] [-w winsize] inimg outimg\n")
- fmt.Fprintf(os.Stderr, "Wipes the sections of an image which are outside the content area.\n")
- flag.PrintDefaults()
- }
- min := flag.Int("m", 30, "Minimum percentage of the image width for the content width calculation to be considered valid.")
- thresh := flag.Float64("t", 0.05, "Threshold for the proportion of black pixels below which a window is determined to be the edge. Higher means more aggressive wiping.")
- wsize := flag.Int("w", 5, "Window size for mask finding algorithm.")
- flag.Parse()
- if flag.NArg() < 2 {
- flag.Usage()
- os.Exit(1)
- }
-
- 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)
-
- clean := preproc.Wipe(gray, *wsize, *thresh, *min)
-
- f, err = os.Create(flag.Arg(1))
- if err != nil {
- log.Fatalf("Could not create file %s: %v\n", flag.Arg(1), err)
- }
- defer f.Close()
- err = png.Encode(f, clean)
- if err != nil {
- log.Fatalf("Could not encode image: %v\n", err)
- }
-}
generated by cgit v1.2.1-24-ge1ad (git 2.21.0) at 2025-10-31 11:13:45 +0000