python处理DICOM并计算三维模型体积

在已知DICOM和三维模型对应掩膜的情况下，计算三维模型的体积。

思路：

1、计算每个体素的体积。每个体素为长方体，x，y为PixelSpacing，z为层间距

使用pydicom.read_file读取DICOM文件，dcm_tag.PixelSpacing获取像素间距，dcm_tag.SliceLocation 获取层间距

2、计算体素的个数

代码如下：

from PIL import Image
import numpy as np
import pydicom
import os

def get_pixels_No(bmp_data_dir):
 pixels_No = 0
 bmp_files = os.listdir(bmp_data_dir)
 for bmp in bmp_files:
   bmp_file = os.path.join(bmp_data_dir,bmp)
   img = Image.open(bmp_file)
   img_array = np.array(img)
   # img_array.dtype为布尔类型，需要转换为Int类型，其累加和恰好为体素总和
   img_array_int = img_array.astype(int)
   pixels_No = pixels_No+img_array_int.sum()
 return pixels_No

def get_pixel_info(dcm_data_dir):

pixel_infos = []
 dcm_files = os.listdir(dcm_data_dir)

dcm_file_1 = os.path.join(dcm_data_dir,dcm_files[0])
 dcm_tag_1 = pydicom.read_file(dcm_file_1)
 # 获取像素间距.
 spacex, spacey = dcm_tag_1.PixelSpacing
 # 获取层间距
 # 有些 dcm图像并不是按照InstanceNumber进行排序的，不能直接用最后一张的slicelocation减去第一张，再除以张数
 SliceLocations = []
 ImagePositon_z = []
 for dcm in dcm_files:
   dcm_file = os.path.join(dcm_data_dir, dcm)
   dcm_tag = pydicom.read_file(dcm_file)
   SliceLocations.append(dcm_tag.SliceLocation)
   ImagePositon_z.append(dcm_tag.ImagePositionPatient[2])
 SliceLocations_max =max(SliceLocations)
 SliceLocations_min =min(SliceLocations)
 ImagePositon_z_max = max(ImagePositon_z)
 ImagePositon_z_min = min(ImagePositon_z)
 print(SliceLocations_max)
 print(SliceLocations_min)
 print(ImagePositon_z_max)
 print(ImagePositon_z_min)
 if SliceLocations_max - SliceLocations_min < 1e-10:
   spacez = abs(ImagePositon_z_max - ImagePositon_z_min)/(len(dcm_files)-1)
 else:
   spacez = abs(SliceLocations_max - SliceLocations_min)/(len(dcm_files)-1)
 pixel_infos = [spacex, spacey, spacez]

return pixel_infos

def get_volume(dcm_data_dir,bmp_data_dir):
 pixel_infos = get_pixel_info(dcm_data_dir)
 pixels_No = get_pixels_No(bmp_data_dir)
 volume=pixel_infos[0]*pixel_infos[1]*pixel_infos[2]*pixels_No/1000
 return volume

# dcm = pydicom.read_file(r"E:\20181210090945_LENG HONGYING F-44Y\Venous\0000.dcm")
# print(dcm)
# print(dcm.ImagePositionPatient[2])
# print(dcm[0x0020, 0x0032].keyword,dcm[0x0020, 0x0032].value)

volume=get_volume(r"E:\20181210090945_LENG HONGYING F-44Y\Venous",r"E:\20181210090945_LENG HONGYING F-44Y\Results\LungL")
print("体积为％.1f"％volume)

来源：https://blog.csdn.net/yaoliuwei1426/article/details/85680152