嗨，我正在尝试开发一个代码（如下），以使用轮廓检测汽车，但问题是我的图像质量太低，轮廓显得稀疏，无法让我检测到使用近似值的汽车，我只是附加了原始图像的链接。 你有什么建议？
import numpy as np
import cv2 as cv
img2=cv.imread('9.bmp')
from PIL import Image
Image.open("9.bmp").save("sample1.jpg")
img2=cv.imread('sample1.jpg')
img = cv.detailEnhance(img2, sigma_s=300, sigma_r=0.10)
imggray=cv.cvtColor(img,cv.COLOR_BGR2GRAY)
blurred = cv.GaussianBlur(imggray, (1,1), 0)
edged = cv.Canny(blurred, 999, 3,True)
_,thresh3 = cv.threshold(edged,254,255,cv.THRESH_TRUNC)
contours,_=cv.findContours(thresh3,cv.RETR_TREE,cv.CHAIN_APPROX_SIMPLE)
cv.drawContours(img,contours,-1,(0,255,0),1)
#for contour in contours:
#
# approx=cv.approxPolyDP(contour,0.0001*cv.arcLength(contour,True),True)
# (x, y, w, h) = cv.boundingRect(approx)
# area = cv.contourArea(contour)
# ar=float(w)/h
# if area<=600 and float(w)<100 and float(w)>5 and ar<=10:
# cv.drawContours(img,[approx],-1,(0,0,255),1)
