白色洗衣机图像划痕检测：现有代码失效的解决方案咨询

问题描述 #

我有一组包含浅、深划痕的白色洗衣机图像数据集，用从StackOverflow拿到的代码检测时输出黑屏，完全识别不出划痕。现有两张测试图：

1. 侧边裁剪并旋转的图像：
2. 侧边裁剪未旋转的图像：

原代码如下：

import cv2
import numpy as np

# load image
img = cv2.imread('./detect-01.jpg')

# convert to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# adaptive threshold 
thresh = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 11, -35)

# apply morphology
kernel = np.ones((3,30),np.uint8)
morph = cv2.morphologyEx(thresh, cv2.MORPH_CLOSE, kernel)
kernel = np.ones((3,35),np.uint8)
morph = cv2.morphologyEx(morph, cv2.MORPH_OPEN, kernel)

# get hough line segments
threshold = 25
minLineLength = 10
maxLineGap = 20
lines = cv2.HoughLinesP(morph, 1, 30*np.pi/360, threshold, minLineLength, maxLineGap)

# draw lines
linear1 = np.zeros_like(thresh)
linear2 = img.copy()
for [line] in lines:
    x1 = line[0]
    y1 = line[1]
    x2 = line[2]
    y2 = line[3]
    cv2.line(linear1, (x1,y1), (x2,y2), 255, 1)
    cv2.line(linear2, (x1,y1), (x2,y2), (0,0,255), 1)

print('number of lines:',len(lines))

# save resulting masked image
cv2.imwrite('scratches_thresh.jpg', thresh)
cv2.imwrite('scratches_morph.jpg', morph)
cv2.imwrite('scratches_lines1.jpg', linear1)
cv2.imwrite('scratches_lines2.jpg', linear2)

# display result
cv2.imshow("thresh", thresh)
cv2.imshow("morph", morph)
cv2.imshow("lines1", linear1)
cv2.imshow("lines2", linear2)
cv2.waitKey(0)
cv2.destroyAllWindows()

原代码修改方案 #

原代码黑屏的核心问题是自适应阈值参数不匹配，加上形态学操作的kernel尺寸过大，直接把划痕特征抹掉了。针对白色洗衣机的划痕，调整如下：

修改后的代码 #

import cv2
import numpy as np

# 加载图像
img = cv2.imread('./detect-01.jpg')
if img is None:
    print("图像加载失败，请检查路径")
    exit()

# 转灰度图
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# 高斯模糊降噪（针对浅划痕）
blur = cv2.GaussianBlur(gray, (3,3), 0)

# 调整自适应阈值：白色背景下，划痕是深色，用THRESH_BINARY_INV反转阈值
thresh = cv2.adaptiveThreshold(blur, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
                               cv2.THRESH_BINARY_INV, 21, 8)

# 形态学操作：用小kernel增强划痕，避免过度腐蚀
kernel = np.ones((2,2), np.uint8)
morph = cv2.morphologyEx(thresh, cv2.MORPH_CLOSE, kernel, iterations=1)
morph = cv2.morphologyEx(morph, cv2.MORPH_OPEN, kernel, iterations=1)

# 霍夫线检测调整参数：适配划痕的长度和角度
threshold = 15
minLineLength = 20
maxLineGap = 5
lines = cv2.HoughLinesP(morph, 1, np.pi/180, threshold, minLineLength, maxLineGap)

# 绘制检测结果
linear1 = np.zeros_like(thresh)
linear2 = img.copy()
if lines is not None:
    print('检测到的线条数:', len(lines))
    for [line] in lines:
        x1, y1, x2, y2 = line
        cv2.line(linear1, (x1,y1), (x2,y2), 255, 1)
        cv2.line(linear2, (x1,y1), (x2,y2), (0,0,255), 2)
else:
    print('未检测到线条')

# 保存结果
cv2.imwrite('scratches_thresh.jpg', thresh)
cv2.imwrite('scratches_morph.jpg', morph)
cv2.imwrite('scratches_lines1.jpg', linear1)
cv2.imwrite('scratches_lines2.jpg', linear2)

# 显示窗口
cv2.imshow("gray", gray)
cv2.imshow("thresh", thresh)
cv2.imshow("morph", morph)
cv2.imshow("lines", linear2)
cv2.waitKey(0)
cv2.destroyAllWindows()

修改要点说明 #

• 增加高斯模糊：过滤图像中的微小噪声，避免误检测，同时突出浅划痕的边缘。
• 反转阈值类型：白色背景下划痕是深色，用THRESH_BINARY_INV让划痕变成白色，背景黑色，符合后续检测逻辑。
• 调整阈值参数：把blockSize改成21，C值改成8，适配洗衣机表面的明暗差异，能更好捕捉浅划痕。
• 缩小形态学kernel：原代码的大尺寸kernel会直接抹平细长划痕，换成(2,2)小kernel，只做轻微的闭开操作修复划痕断点。
• 优化霍夫线参数：降低阈值、调整最小线长和最大间隙，适配划痕的细长特征，避免检测无关短线条。
• 增加图像加载判断：避免因路径错误导致后续崩溃。

替代方案：边缘检测+轮廓提取 #

如果霍夫线检测效果仍不理想，可尝试用Canny边缘检测结合轮廓提取的方案，更适合不规则划痕：

import cv2
import numpy as np

img = cv2.imread('./detect-01.jpg')
if img is None:
    print("图像加载失败")
    exit()

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
blur = cv2.GaussianBlur(gray, (5,5), 0)

# Canny边缘检测，调整高低阈值适配划痕
edges = cv2.Canny(blur, 30, 150)

# 形态学膨胀增强边缘
kernel = np.ones((2,2), np.uint8)
edges_dilated = cv2.dilate(edges, kernel, iterations=1)

# 提取轮廓
contours, _ = cv2.findContours(edges_dilated, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

# 过滤掉过小的轮廓，只保留疑似划痕的区域
result = img.copy()
min_contour_length = 15
for cnt in contours:
    if cv2.arcLength(cnt, False) > min_contour_length:
        cv2.drawContours(result, [cnt], 0, (0,0,255), 1)

cv2.imwrite('scratches_contours.jpg', result)
cv2.imshow("contours", result)
cv2.waitKey(0)
cv2.destroyAllWindows()

方案优势 #

• 对不规则、非直线的划痕适配性更强，比如弯曲的浅划痕。
• 轮廓过滤可以有效排除灰尘、污渍等小干扰。

内容的提问来源于stack exchange，提问作者Lean Learner