[es] - Crop image aplikacija

Panta_
Aleksandar Pantić
Kragujevac

Član broj: 214959
Poruke: 790

+162 Profil

^{12.10.2022. u 16:18 - pre 18 meseci}

Citat:

Da li bi mogla neka skripta da se napravi?
I kako je ubaciti? I gde?

Slično se traži na onom linku slike koji si postavio u prvoj poruci, a ne Windows aplikacija koja u slici traži konture određenog oblika i seče sliku do istih. Kao što je navedeno na pomenutom linku, to možeš da odradiš pomoću OpenCV API-ja koji je dostupan za C++, Java i Python programske jezike.

Ja sam se pomalo igrao sa pomenutom bibliotekom u Python-u i dobio ne tako loše rezultate.

Code (python):

import pathlib

import sys

import cv2 as cv

import numpy as np

def read_image(image):

    img = cv.imread(image)

    # if img.shape[1] > 1000:

    #     img = cv.resize(img, (img.shape[1]//2, img.shape[0]//2),

    #                     interpolation=cv.INTER_CUBIC)

    return img

def get_corners(image):

    gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)

    # blur = cv.GaussianBlur(gray, (5,5), cv.BORDER_DEFAULT)

    # canny = cv.Canny(blur, 125, 175)

    # cv.imshow('Canny', canny)

    _, thresh = cv.threshold(gray, 125, 255, cv.THRESH_BINARY)

    # cv.imshow('Thresh', thresh)

    # https://docs.opencv.org/4.x/dd...orial_py_contour_features.html

    contours, _ = cv.findContours(thresh, cv.RETR_LIST,

                                  cv.CHAIN_APPROX_SIMPLE)  # cv.CHAIN_APPROX_NONE

    contours = sorted(contours, key=cv.contourArea, reverse=True)

    for contour in contours:

        # approximate the contour

        perimeter = 0.05 * cv.arcLength(contour, True)

        corners = cv.approxPolyDP(contour, perimeter, True)

        # if contour with 4 points founded, break the loop

        if len(corners) == 4:

            break

    return corners

# Copy/Paste from

# https://pyimagesearch.com/2014...perspective-transform-example/

# or install: https://pypi.org/project/imutils/

def order_points(pts):

     # initialzie a list of coordinates that will be ordered

     # such that the first entry in the list is the top-left,

     # the second entry is the top-right, the third is the

     # bottom-right, and the fourth is the bottom-left

     rect = np.zeros((4, 2), dtype = "float32")

     # the top-left point will have the smallest sum, whereas

     # the bottom-right point will have the largest sum

     s = pts.sum(axis = 1)

     rect[0] = pts[np.argmin(s)]

     rect[2] = pts[np.argmax(s)]

     # now, compute the difference between the points, the

     # top-right point will have the smallest difference,

     # whereas the bottom-left will have the largest difference

     diff = np.diff(pts, axis = 1)

     rect[1] = pts[np.argmin(diff)]

     rect[3] = pts[np.argmax(diff)]

     # return the ordered coordinates

     return rect

def four_point_transform(image, pts):

     # obtain a consistent order of the points and unpack them

     # individually

     rect = order_points(pts)

     (tl, tr, br, bl) = rect

     # compute the width of the new image, which will be the

     # maximum distance between bottom-right and bottom-left

     # x-coordiates or the top-right and top-left x-coordinates

     widthA = np.sqrt(((br[0] - bl[0]) ** 2) + ((br[1] - bl[1]) ** 2))

     widthB = np.sqrt(((tr[0] - tl[0]) ** 2) + ((tr[1] - tl[1]) ** 2))

     maxWidth = max(int(widthA), int(widthB))

     # compute the height of the new image, which will be the

     # maximum distance between the top-right and bottom-right

     # y-coordinates or the top-left and bottom-left y-coordinates

     heightA = np.sqrt(((tr[0] - br[0]) ** 2) + ((tr[1] - br[1]) ** 2))

     heightB = np.sqrt(((tl[0] - bl[0]) ** 2) + ((tl[1] - bl[1]) ** 2))

     maxHeight = max(int(heightA), int(heightB))

     # now that we have the dimensions of the new image, construct

     # the set of destination points to obtain a "birds eye view",

     # (i.e. top-down view) of the image, again specifying points

     # in the top-left, top-right, bottom-right, and bottom-left

     # order

     dst = np.array([

          [0, 0],

          [maxWidth - 1, 0],

          [maxWidth - 1, maxHeight - 1],

          [0, maxHeight - 1]], dtype = "float32")

     # compute the perspective transform matrix and then apply it

     M = cv.getPerspectiveTransform(rect, dst)

     warped = cv.warpPerspective(image, M, (maxWidth, maxHeight))

     # return the warped image

     return warped

if __name__ == '__main__':

    if len(sys.argv) > 1:

        img_path = pathlib.Path(sys.argv[1])

    else:

        img_path = pathlib.Path(__file__).parent

    output_dir = f'{img_path.as_posix()}/output'

    pathlib.Path(output_dir).mkdir(exist_ok=True)

    for img in img_path.iterdir():

        if img.suffix.lower() in ('.png', '.jpg', '.jpeg'):

            image = read_image(img.as_posix())

            corners = get_corners(image)

            new_image = four_point_transform(image, corners.reshape(4, 2))

            image_name = f'{img.name.removesuffix(img.suffix)}_cropped{img.suffix}'

            cv.imwrite(f'{output_dir}/{image_name}', new_image)

Za pokretanje Python skripta je potrbno preuzeti i instalirati Python, zatim instalirati opencv-python paket. Python script pokrenuti sa python ime_skripta.py ako se slike nalaze u istom direktorijumu kao i sam skript ili sa python ime_skripta.py "/putanja/do/direktorijuma/sa/slikama".

Prikačeni fajlovi

primer_1.png - 1633.25k

primer_2.png - 261.92k

primer_3.png - 1006.29k

primer_4.png - 361.57k

crop_images.py - 4.15k

Odgovor na temu