7
3

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 5 years have passed since last update.

OpenCVで主人公レンズ

Last updated at Posted at 2017-12-17

主人公レンズとは

@itopoid さんが作ったノベルゲーム風フレームのグッズ。透明なプラスチック製?でゲームのセリフ画面をリアルに持ってこれる。

DPOdG2KUEAAiCCO.jpg
https://twitter.com/itopoid/status/933262577651433478

Web版を作ろうと思ったらもうあった。でもめげずにOpenCVで実装してみる。

素材集め

フリー素材サイトから集める。

枠は空想曲線より、サンプル背景はfreephotos.ccより取得。

枠:
message_box.png

背景:
hallway.jpg

テキストから画像を生成

OpenCVでやろうと思ったが日本語テキストは単体だと無理っぽい。しかたがないのでPillowを使う。

main.py
from PIL import Image
from PIL import ImageDraw
from PIL import ImageFont

def draw_text(img, text, color):
  draw = ImageDraw.Draw(img)
  draw.font = ImageFont.truetype("/Library/Fonts/ヒラギノ丸ゴ ProN W4.ttc", 20)

  position = np.array([0, 0])
  draw.text(position, text, color)

img = Image.new("RGBA", (750, 80))
text = "いったいアイツはどこにいったんだ。まだOpenCVのビルドの途中だってのに、、、"
draw_text(img, text, (38, 57, 54))
img.save('output/text.png')

img = Image.new("RGBA", (250, 30))
text = "CTO"
draw_text(img, text, (255, 255, 255))
img.save('output/speaker.png')

OpenCVで画像合成

alphaチャンネル付きで読み込んで合成する。imreadに-1のオプションをつけると計4チャンネルで読み込まれる

main.py
def merge_images(background, foreground_with_alpha, o_h, o_w):
    alpha = foreground_with_alpha[:,:,3]
    alpha = cv2.cvtColor(alpha, cv2.COLOR_GRAY2BGR) # 1 -> 3
    alpha = alpha / 255.0

    foreground = foreground_with_alpha[:,:,:3]

    f_h, f_w, _ = foreground.shape
    b_h, _, _ = background.shape
    background[b_h-f_h-o_h:b_h-o_h, o_w:f_w+o_w] = (background[b_h-f_h-o_h:b_h-o_h, o_w:f_w+o_w] * (1.0 - alpha)).astype('uint8')
    background[b_h-f_h-o_h:b_h-o_h, o_w:f_w+o_w] = (background[b_h-f_h-o_h:b_h-o_h, o_w:f_w+o_w] + (foreground * alpha)).astype('uint8')

    return background

 # -1 to read alpha channel
# foreground: (210, 960, 4)
message_box_with_alpha =cv2.imread('samples/message_box.png', -1)
# background: (640, 960, 3)
hallway = cv2.imread('samples/hallway.jpg')
result = merge_images(hallway, message_box_with_alpha, 0, 0)

# text (80, 750, 3)
text_with_alpha = cv2.imread('output/text.png', -1)
result = merge_images(result, text_with_alpha, 40, 105)

# speaker (30, 250, 3)
speaker_with_alpha = cv2.imread('output/speaker.png', -1)
result = merge_images(result, speaker_with_alpha, 145, 50)

cv2.imwrite("output/result.png", result)

結果

result.png

ソースコード

参考

7
3
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
7
3

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?