一组图像中颜色(hue)值(0-359)出现的总和 [英] Sum of the occurence of color(hue) values(0-359) from a set of images

问题描述

我有一个装满图像的文件夹，我想找到出现次数最少的色相值.为此，我为所有色相值创建了一个长度为360的数组，拍摄了我文件夹中的所有图像，进行了遍历，对于每个像素，我在代表色相值的索引处为数组添加了+1.例如，如果我的像素中的色相值为0，则在数组中的索引0处添加+1. 我的问题是:有没有更快或更有效的方法?

I have a folder full of images and i want to find the Hue Values with the smallest occurence. For that i create an array with length 360 for all the hue values, take all the images in my folder, go through it and for each pixel i add +1 in my array at the index which represent the hue value. If i have for example the hue value 0 in my pixel, i add +1 in my array at index 0. My question is: is there a faster or more efficent way to do that?

这是我的代码:

path = 'path'
sub_path = 'sub_path'
sumHueOcc = np.zeros((360, 1), dtype=np.uint64)

for item in dirs:
    fullpath = os.path.join(path,item)
    pathos = os.path.join(sub_path,item)
    if os.path.isfile(fullpath):
        f, e = os.path.splitext(pathos)
        img = np.array(Image.open(fullpath))
        img = np.float32(img)     
        imgHSV = cv2.cvtColor(img, cv2.COLOR_RGB2HSV_FULL) #RGB because numpy RGB

        # want to work with hue only
        h, s, v = cv2.split(imgHSV)

        # the hue values in one large array
        Z = np.array(h, copy=True)
        Z = Z.reshape((-1, 1))

        # convert to np.float32
        Z = np.uint64(Z)

        # count each appearence from hue values
        for z in Z:
            sumHueOcc[z] = sumHueOcc[z] + 1


max = np.argmax(sumHueOcc)
min = np.argmin(sumHueOcc)
print("Minimum 1")
print(min)
sumHueOcc[min] += max
min = np.argmin(sumHueOcc)
print("Minimum 2")
print(min)
sumHueOcc[min] += max
min = np.argmin(sumHueOcc)
print("Minimum 3")
print(min)
sumHueOcc[min] += max
min = np.argmin(sumHueOcc)
print("Minimum 4")
print(min)

推荐答案

我们可以使用np.bincount进行计数.

We can use np.bincount to do the counting.

因此，我们从int64开始对输出数组进行初始化-

So, we initialize the output array at the start with int64 -

sumHueOcc_out = np.zeros((180, 1), dtype=np.int64) 

然后，在循环内部，我们替换涉及循环的最里面的部分-

Then, inside the loops, we replace the innermost section involving loops -

# the hue values in one large array
Z = np.array(h, copy=True)
Z = Z.reshape((-1, 1))

# convert to np.float32
Z = np.uint64(Z)

# count each appearence from hue values
for z in Z:
    sumHueOcc[z] = sumHueOcc[z] + 1

具有bincount替代-

sumHueOcc_out.flat += np.bincount(h.astype(np.int64).ravel(),minlength=sumHueOcc.size)

这篇关于一组图像中颜色(hue)值(0-359)出现的总和的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！