如何裁剪无人机拍摄到的太阳能电池板？

Question

我目前正在研究从无人机拍摄的图像中裁剪太阳能电池板（附上示例图像）。我尝试过使用轮廓，但没有得到正确的结果。它没有检测到图像中的所有太阳能电池板，其中一些电池板丢失了。我自己击中了这里。我该如何进一步进行？请帮我解决这个问题。

谢谢你，

示例代码：

import cv2
import numpy as np
img = cv2.imread('D:\\SolarPanel Images\\solarpanel.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
blur = cv2.GaussianBlur(gray,(5,5),0)
edges = cv2.Canny(blur,100,200)    
th3 = cv2.adaptiveThreshold(edges,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C,cv2.THRESH_BINARY,11,2)

im2, contours, hierarchy = cv2.findContours(th3, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
print("Len of contours",len(contours)
try: hierarchy = hierarchy[0]
except: hierarchy = []

height, width,  = edges.shape
min_x, min_y = width, height
max_x = max_y = 0

# computes the bounding box for the contour, and draws it on the image,
for contour, hier in zip(contours, hierarchy):
    area = cv2.contourArea(contour)

    if area > 10000 and area < 250000:
        (x,y,w,h) = cv2.boundingRect(contour)
        min_x, max_x = min(x, min_x), max(x+w, max_x)
        min_y, max_y = min(y, min_y), max(y+h, max_y)
        if w > 80 and h > 80:
            cv2.rectangle(img, (x,y), (x+w,y+h), (255, 0, 0), 2)

            cv2.imshow('cont imge', img)
            cv2.waitKey(0)

Answer 1

要在图像中找到重要对象与背景清晰可辨的轮廓，您始终可以尝试将图像转换为 HSV 格式，然后再转换轮廓。我做了以下事情：

import cv2
import numpy as np
img = cv2.imread('panel.jpg')
hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV) 
ret,thresh1 = cv2.threshold(hsv[:,:,0],100,255,cv2.THRESH_BINARY)
im2, contours, hierarchy = cv2.findContours(thresh1, cv2.RETR_TREE, 
cv2.CHAIN_APPROX_SIMPLE)
try: hierarchy = hierarchy[0]
except: hierarchy = []
for contour, hier in zip(contours, hierarchy):
    area = cv2.contourArea(contour)
    if area > 10000 and area < 250000:
       rect = cv2.minAreaRect(contour)
       box = cv2.boxPoints(rect)
       box = np.int0(box)
       cv2.drawContours(img,[box],0,(0,0,255),2)
       cv2.imshow('cont imge', img)
       cv2.waitKey(0)
cv2.imwrite("result.jpg",img)

结果：

Answer 2

您似乎面临着无人机图像中“太阳能电池板”裁剪的轮廓检测挑战。考虑通过应用直方图均衡、降噪和边缘增强等技术来改进图像预处理。此外，探索基于机器学习的目标检测方法（例如 YOLO 或 Mask R-CNN）可能会产生更好的结果。