WIN10下Anaconda安裝 環境變量

2. 安裝Anaconda3-4.3.1-Windows-x86_64，默認Python版本為3.6

3. 安裝完以後，打開Anaconda Prompt，輸入清華的倉庫鏡像，更新包更快：

conda config --add channels/anaconda/pkgs/free/

conda config --set show_channel_urls yes

4. 建立TensorFlow空間：conda create -n tensorflow python=3.5，設置Python版本為3.5

5. 激活TensorFlow空間：activate tensorflow

6. 安裝TensorFlow：pip install--ignore-installed --upgrade tensorflow_gpu-1.0.0-cp35-cp35m-win_amd64.whl

7. 測試TensorFlow：

import tensorflow as tf

hello = tf.constant('Hello, TensorFlow!')

sess = tf.Session()

print(sess.run(hello))

設置SSD運行環境

1． 安裝numpy（anaconda 離線安裝）：pipinstall numpy-1.12.1-cp35-none-win_amd64.whl

2． 安裝matplotlib（anaconda 離線安裝）：pipinstall matplotlib-2.0.1-cp35-cp35m-win_amd64.whl

3． 安裝opencv（離線）：anaconda中安裝：pipinstall opencv_python-3.2.0+contrib-cp35-cp35m-win_amd64.whl

4． 下載TensorFlow版本的SSD：git clone /balancap/SSD-Tensorflow.git，或者下載壓縮包解壓

5． 解壓TensorFlowssd目錄下的/checkpoint裏的ssd_300_vgg.ckpt.zip，得到模型參數

6． 安裝pycharm-community-2017.1.2.exe，python編輯器，File->Setting->Project:Python->ProjectInterpreter:修改Python版本到TensorFlow工作空間下的python

7． 在notebook下新建工程，新建test_ssd.py文件

8． 在pycharm中打開ssd_notebook.ipynb，復制非註釋的內容至test_ssd.py下

9． 修改test_ssd.py：

# Test on some demoimage and visualize output.

#path = '../demo/'

#image_names = sorted(os.listdir(path))

#print(image_names)

#for it in image_names:

cam=cv2.VideoCapture(0)

success, img = cam.read()

while success:

#img = cv2.imread(path+it)#mpimg.imread(path + it)

t1=cv2.getTickCount()

rclasses, rscores, rbboxes = process_image(img)

visualization.bboxes_draw_on_img(img,rclasses, rscores, rbboxes, visualization.colors_plasma)

t2=cv2.getTickCount()

print('time consumption:%.3f ms'%(1000*(t2-t1)/cv2.getTickFrequency()))

cv2.imshow('test',img)

c=cv2.waitKey(1)

if c==27:

break

# visualization.bboxes_draw_on_img(img, rclasses,rscores, rbboxes, visualization.colors_plasma)

#visualization.plt_bboxes(img, rclasses,rscores, rbboxes)

success, img = cam.read()

10． 運行程序test_ssd.py