首頁 > 軟體

OpenCV實戰案例之車道線識別詳解

2022-10-18 14:01:13

一、首先進行canny邊緣檢測,為獲取車道線邊緣做準備

import cv2
 
gray_img = cv2.imread('img.jpg',cv2.IMREAD_GRAYSCALE)
canny_img = cv2.Canny(gray_img,50,100)
cv2.imwrite('canny_img.jpg',canny_img)
cv2.imshow('canny',canny_img)
 
cv2.waitKey(0)

二、進行ROI提取獲取確切的車道線邊緣(紅色線內部)

方法:在影象中,黑色表示0,白色為1,那麼要保留矩形內的白色線,就使用邏輯與,當然前提是影象矩形外也是0,那麼就採用建立一個全0影象,然後在矩形內全1,之後與之前的canny影象進行與操作,即可得到需要的車道線邊緣。

import cv2
import numpy as np
 
canny_img = cv2.imread('canny_img.jpg',cv2.IMREAD_GRAYSCALE)
roi = np.zeros_like(canny_img)
roi = cv2.fillPoly(roi,np.array([[[0, 368],[300, 210], [340, 210], [640, 368]]]),color=255)
roi_img = cv2.bitwise_and(canny_img, roi)
cv2.imwrite('roi_img.jpg',roi_img)
cv2.imshow('roi_img',roi_img)
cv2.waitKey(0)

三、利用概率霍夫變換獲取直線,並將斜率正數和複數的線段給分割開來

TIPs:使用霍夫變換需要將影象先二值化

概率霍夫變換函數:

  • lines=cv2.HoughLinesP(image, rho,theta,threshold,minLineLength, maxLineGap)
  • image:影象,必須是8位元單通道二值影象
  • rho:以畫素為單位的距離r的精度,一般情況下是使用1
  • theta:表示搜尋可能的角度,使用的精度是np.pi/180
  • threshold:閾值,該值越小,判定的直線越多,相反則直線越少
  • minLineLength:預設為0,控制接受直線的最小長度
  • maxLineGap:控制接受共線線段的最小間隔,如果兩點間隔超過了引數,就認為兩點不在同一直線上,預設為0
  • lines:返回值由numpy.ndarray構成,每一對都是一對浮點數,表示線段的兩個端點
import cv2
import numpy as np
 
#計算斜率
def calculate_slope(line):
    x_1, y_1, x_2, y_2 = line[0]
    return (y_2 - y_1) / (x_2 - x_1)
 
edge_img = cv2.imread('masked_edge_img.jpg', cv2.IMREAD_GRAYSCALE)
#霍夫變換獲取所有線段
lines = cv2.HoughLinesP(edge_img, 1, np.pi / 180, 15, minLineLength=40,
                        maxLineGap=20)
 
#利用斜率劃分線段
left_lines = [line for line in lines if calculate_slope(line) < 0]
right_lines = [line for line in lines if calculate_slope(line) > 0]

四、離群值過濾,剔除斜率相差過大的線段

流程:

  • 獲取所有的線段的斜率,然後計算斜率的平均值
  • 遍歷所有斜率,計算和平均斜率的差值,尋找最大的那個斜率對應的直線,如果差值大於閾值,那麼就從列表中剔除對應的線段和斜率
  • 迴圈執行操作,直到剩下的全部都是小於閾值的線段
def reject_abnormal_lines(lines, threshold):
    slopes = [calculate_slope(line) for line in lines]
    while len(lines) > 0:
        mean = np.mean(slopes)
        diff = [abs(s - mean) for s in slopes]
        idx = np.argmax(diff)
        if diff[idx] > threshold:
            slopes.pop(idx)
            lines.pop(idx)
        else:
            break
    return lines
 
reject_abnormal_lines(left_lines, threshold=0.2)
reject_abnormal_lines(right_lines, threshold=0.2)

五、最小二乘擬合,實現將左邊和右邊的線段互相擬合成一條直線,形成車道線

流程:

  • 取出所有的直線的x和y座標,組成列表,利用np.ravel進行將高維轉一維陣列
  • 利用np.polyfit進行直線的擬合,最終得到擬合後的直線的斜率和截距,類似y=kx+b的(k,b)
  • 最終要返回(x_min,y_min,x_max,y_max)的一個np.array的資料,那麼就是用np.polyval求多項式的值,舉個example,np.polyval([3,0,1], 5) # 3 * 5**2 + 0 * 5**1 + 1,即可以獲得對應x座標的y座標。
def least_squares_fit(lines):
    # 1. 取出所有座標點
    x_coords = np.ravel([[line[0][0], line[0][2]] for line in lines])
    y_coords = np.ravel([[line[0][1], line[0][3]] for line in lines])
 
    # 2. 進行直線擬合.得到多項式係數
    poly = np.polyfit(x_coords, y_coords, deg=1)
    print(poly)
    # 3. 根據多項式係數,計算兩個直線上的點,用於唯一確定這條直線
    point_min = (np.min(x_coords), np.polyval(poly, np.min(x_coords)))
    point_max = (np.max(x_coords), np.polyval(poly, np.max(x_coords)))
    return np.array([point_min, point_max], dtype=np.int)
 
print("left lane")
print(least_squares_fit(left_lines))
print("right lane")
print(least_squares_fit(right_lines))

六、繪製線段

cv2.line(img, tuple(left_line[0]), tuple(left_line[1]), color=(0, 255, 255), thickness=5)
cv2.line(img, tuple(right_line[0]), tuple(right_line[1]), color=(0, 255, 255), thickness=5)

全部程式碼(視訊顯示)

import cv2
import numpy as np
 
def get_edge_img(color_img, gaussian_ksize=5, gaussian_sigmax=1,
                 canny_threshold1=50, canny_threshold2=100):
    """
    灰度化,模糊,canny變換,提取邊緣
    :param color_img: 彩色圖,channels=3
    """
    gaussian = cv2.GaussianBlur(color_img, (gaussian_ksize, gaussian_ksize),
                                gaussian_sigmax)
    gray_img = cv2.cvtColor(gaussian, cv2.COLOR_BGR2GRAY)
    edges_img = cv2.Canny(gray_img, canny_threshold1, canny_threshold2)
    return edges_img
 
def roi_mask(gray_img):
    """
    對gray_img進行掩膜
    :param gray_img: 灰度圖,channels=1
    """
    poly_pts = np.array([[[0, 368], [300, 210], [340, 210], [640, 368]]])
    mask = np.zeros_like(gray_img)
    mask = cv2.fillPoly(mask, pts=poly_pts, color=255)
    img_mask = cv2.bitwise_and(gray_img, mask)
    return img_mask
 
 
def get_lines(edge_img):
    """
    獲取edge_img中的所有線段
    :param edge_img: 標記邊緣的灰度圖
    """
 
    def calculate_slope(line):
        """
        計算線段line的斜率
        :param line: np.array([[x_1, y_1, x_2, y_2]])
        :return:
        """
        x_1, y_1, x_2, y_2 = line[0]
        return (y_2 - y_1) / (x_2 - x_1)
 
    def reject_abnormal_lines(lines, threshold=0.2):
        """
        剔除斜率不一致的線段
        :param lines: 線段集合, [np.array([[x_1, y_1, x_2, y_2]]),np.array([[x_1, y_1, x_2, y_2]]),...,np.array([[x_1, y_1, x_2, y_2]])]
        """
        slopes = [calculate_slope(line) for line in lines]
        while len(lines) > 0:
            mean = np.mean(slopes)
            diff = [abs(s - mean) for s in slopes]
            idx = np.argmax(diff)
            if diff[idx] > threshold:
                slopes.pop(idx)
                lines.pop(idx)
            else:
                break
        return lines
 
    def least_squares_fit(lines):
        """
        將lines中的線段擬合成一條線段
        :param lines: 線段集合, [np.array([[x_1, y_1, x_2, y_2]]),np.array([[x_1, y_1, x_2, y_2]]),...,np.array([[x_1, y_1, x_2, y_2]])]
        :return: 線段上的兩點,np.array([[xmin, ymin], [xmax, ymax]])
        """
        x_coords = np.ravel([[line[0][0], line[0][2]] for line in lines])
        y_coords = np.ravel([[line[0][1], line[0][3]] for line in lines])
        poly = np.polyfit(x_coords, y_coords, deg=1)
        point_min = (np.min(x_coords), np.polyval(poly, np.min(x_coords)))
        point_max = (np.max(x_coords), np.polyval(poly, np.max(x_coords)))
        return np.array([point_min, point_max], dtype=np.int)
 
    # 獲取所有線段
    lines = cv2.HoughLinesP(edge_img, 1, np.pi / 180, 15, minLineLength=40,
                            maxLineGap=20)
    # 按照斜率分成車道線
    left_lines = [line for line in lines if calculate_slope(line) > 0]
    right_lines = [line for line in lines if calculate_slope(line) < 0]
    # 剔除離群線段
    left_lines = reject_abnormal_lines(left_lines)
    right_lines = reject_abnormal_lines(right_lines)
 
    return least_squares_fit(left_lines), least_squares_fit(right_lines)
 
def draw_lines(img, lines):
    left_line, right_line = lines
    cv2.line(img, tuple(left_line[0]), tuple(left_line[1]), color=(0, 255, 255),
             thickness=5)
    cv2.line(img, tuple(right_line[0]), tuple(right_line[1]),
             color=(0, 255, 255), thickness=5)
 
def show_lane(color_img):
    edge_img = get_edge_img(color_img)
    mask_gray_img = roi_mask(edge_img)
    lines = get_lines(mask_gray_img)
    draw_lines(color_img, lines)
    return color_img
 
capture = cv2.VideoCapture('video.mp4')
while True:
    ret, frame = capture.read()
    if not ret:
        break
    frame = show_lane(frame)
    cv2.imshow('frame', frame)
    cv2.waitKey(10)

總結

到此這篇關於OpenCV實戰案例之車道線識別詳解的文章就介紹到這了,更多相關OpenCV車道線識別內容請搜尋it145.com以前的文章或繼續瀏覽下面的相關文章希望大家以後多多支援it145.com!


IT145.com E-mail:sddin#qq.com