Question

我一直在尝试使用imread包中的cv2来阅读jpeg图片。到目前为止，除了这种看似随机选择的图像之外，它还是很棒的。

问题在于，由于我将来自同一来源的所有图像都保存在同一个文件夹中，显然imread处理了多个图像而不是其他图像，没有明显的原因。我仔细检查并确保所有图像都工作正常，类型相同等。但是，如下面附带的屏幕截图所示，只有选定的图像返回了合理的结果，而其他所有图像都返回了None。我在这里使用的确切代码是：

img = cv2.imread(image,1) #most None

有人可以建议可能出错的地方吗？非常感谢！

我的代码作为一个整体：

images = os.listdir(os.path.join(input_dir, folder))
os.chdir(os.path.join(input_dir, folder))
#index += 1
index_array = []
out = []

# a quick recursion of summing up nested lists
rec = lambda x: sum(map(rec, x)) if isinstance(x, list) else x

minLineLength = 20 ##? how to set these?
maxLineGap = 5 ##? how to set these?    

for image in images:
    print(image)
    if image == ".DS_Store" or (not image.endswith(".jpeg")):
        continue

    else:
         img = cv2.imread(image,1) 
#cv2.cv.LoadImage(image,CV_LOAD_IMAGE_COLOR)## #Opening image
         if img is None:
            print("None")
            continue
         elif img is not None:
            print("Pppppppppaaaaaaaaaaassssssssss!")
            gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY); 
            edges = cv2.Canny(gray, 50, 120); 
            nedges = rec(edges.tolist())/255.0; 
            lines = cv2.HoughLinesP(edges, 1, np.pi/180, 100, minLineLength, maxLineGap)
            if lines is None:
                nlines = 0
            elif lines is not None:
                nlines = lines.shape[0]

            img_blur = cv2.GaussianBlur(gray, (5, 5), 0)
            cimg = cv2.cvtColor(img_blur, cv2.COLOR_GRAY2BGR)
            circles = cv2.HoughCircles(img_blur, cv2.HOUGH_GRADIENT, 1, 120, param1=100, param2=30, minRadius=0, maxRadius=0)
            if circles is None:
                ncircles = 0
            elif circles is not None:
                circles = np.uint16(np.around(circles))
                ncircles = circles.shape[1]

            gray_blur = cv2.normalize(img_blur, img_blur, alpha=0, beta=255, norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_8UC1)
            ret, thresh = cv2.threshold(gray_blur, 230, 255, cv2.THRESH_BINARY_INV)
            square_cnts = []
            thresh = cv2.morphologyEx(thresh, cv2.MORPH_CLOSE, np.ones((5, 5), np.uint8))
            tmpimage, contours, h = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
            #print("D")
            cnt0 = contours[0]
            tmp_h = h[::-1]
            appr = []
            npoly = []
            for index, cnt in enumerate(contours[::-1]):
                approx = cv2.approxPolyDP(cnt, 0.1*cv2.arcLength(cnt, True), True)
                if approx is None:
                    continue
                elif approx is not None:
                    appr.append(approx)
                    npoly.append(len(approx))
            new = [nedges, nlines, ncircles, sum(npoly)*1.0/len(npoly)]
            out.append(new) #Adding new image to array shape of (x, 3, 100, 100) where x is image number
            index_array.append(image)

imread在某些图像上返回None，但在同一文件夹中不返回其他图像

0 个答案: