视线跟踪对视频进行ROI剪辑

解决sampling rate不同时的处理办法
这里有的数据是 gaze point 机器测得pupil 和gaze数据是60frame/s，而视频本来的帧数是30frame/s
所以要先把视频30frame 改为60frame

from moviepy.editor import *
#把二十四个视频frame =30--》转为frame =60
def video_frame():
    clip = VideoFileClip('E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/stimuli/stimuli/video/'+str(num[n])+'.mp4')

    clip.write_videofile('E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/stimuli/stimuli/video_new/'+str(num[n])+'.mp4',
                         fps=60)
    clip.reader.close()

然后将视频转为图片

#将24个视频转为图片
def video_image():
    vc = cv2.VideoCapture('E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/stimuli/stimuli/video_new/'+str(num[n])+'.mp4')
    c = 1
    if vc.isOpened():
        rval, frame = vc.read()
    else:
        rval = False
    while rval:
        rval, frame = vc.read()
        cv2.imwrite(path_out + str(c) + '.png',
            frame)
        c = c + 1
        cv2.waitKey(1)
    vc.release()

跟随着视线将图片剪切成（100*100）的ROI图片

def ROI_video():
    data_allgaze = pd.DataFrame(pd.read_csv(gaze_path, encoding="utf-8", engine='python'))
    FPOGX = np.array(data_allgaze['FPOGX'])  # data length : 60fps * time (video time)
    FPOGY = np.array(data_allgaze['FPOGY'])
    picture_name = os.listdir(img_path)
    picture_name = natsorted(picture_name)
    print(str(len(FPOGX)) + '_' + str(len(picture_name)))
    if len(FPOGX) >= len(picture_name):

        for i in range(0, len(picture_name)):
            img_rd = cv2.imread(img_path + picture_name[i], cv2.IMREAD_UNCHANGED)
            x1 = 1280 * abs(FPOGX[int(i)])
            y1 = 720 * abs(FPOGY[int(i)])
            # print(int(y1 - 50), int(y1 + 50), int(x1 - 50), int(x1 + 50))
            img = img_rd[int(y1 - 50):int(y1 + 50), int(x1 - 50):int(x1 + 50)]
            filename = path_out + picture_name[i]
            # cv2.imwrite(path_out + picture_name[i], img)
            try:
                ext = os.path.splitext(filename)[1]
                result, n = cv2.imencode(ext, img, params=None)
                if result:
                    with open(filename, mode='w+b') as f:
                        n.tofile(f)

            except Exception as e:
                print(e)



    if len(FPOGX) <= len(picture_name):
        for i in range(0, len(FPOGX)):
            img_rd = cv2.imread(img_path + picture_name[i], cv2.IMREAD_UNCHANGED)
            x1 = 1280 * abs(FPOGX[int(i)])
            y1 = 720 * abs(FPOGY[int(i)])
            # print(int(y1 - 50), int(y1 + 50), int(x1 - 50), int(x1 + 50))
            img = img_rd[int(y1 - 50):int(y1 + 50), int(x1 - 50):int(x1 + 50)]
            filename = path_out + picture_name[i]
            try:
                ext = os.path.splitext(filename)[1]
                print(ext)
                result, n = cv2.imencode(ext, img, params=None)
                if result:
                    with open(filename, mode='w+b') as f:
                        n.tofile(f)

            except Exception as e:
                print(e)
                continue

#

num = [3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24]
users = ['']#被实验者的名字
#        ,
for n in range(len(num)):
    img_path = 'E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/stimuli/stimuli/video_new/image/'+str(num[n])+'/'
    for user in range(len(users)):
        gaze_path = 'E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/GAZE_DATA_backup/191202/'+users[user]+'/data_process/Experimenter/all_gaze/'+str(num[n])+'_all_gaze.csv'

        print(users[user] + str(num[n]))
        path_out = 'E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/GAZE_DATA_backup/191202/'+users[user]+'/data_process/ROI/roi_picture_7_27/'+str(num[n])+'/'
        mkdir(path_out)

        # print(path_out_new)
        ROI_video()

#将图片保存为一秒储存在一个文件夹中

#路径中有韩文的cv2读取图片的代码
def imread(filename, flags=cv2.IMREAD_COLOR, dtype=np.uint8):
    try:
        n = np.fromfile(filename, dtype)
        img = cv2.imdecode(n, flags)
        return img
    except Exception as e:
        print(e)
        return None
 #路径中有韩文的cv2写入图片的代码
def imwrite(filename, img, params=None):
    try:
        ext = os.path.splitext(filename)[1]
        result, n = cv2.imencode(ext, img, params)
        if result:
            with open(filename, mode='w+b') as f:
                n.tofile(f)
            return True
        else:
            return False
    except Exception as e:
        print(e)
    return False
def clip_image():
    # img_path = 'E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/stimuli/stimuli/video_new/1_1/'
    picture_name = os.listdir(img_path)
    picture_name = natsorted(picture_name)
    for i in range(1, len(picture_name)):
        if (i % 60) == 0:
            path_save = path_out + str(i) + '/'
            mkdir(path_save)
            for j in range(int(i - 60), int(i)):
                print(j)
                #当路径中有韩文存在的时候cv2读取和写入图片的代码
                img = imread(img_path + picture_name[j],flags=cv2.IMREAD_COLOR, dtype=np.uint8)
                imwrite(path_save + picture_name[j], img,params=None)
                #当路径都是英文的时候的代码
                # img = cv2.imread(img_path + picture_name[j])
                # cv2.imwrite(path_save + picture_name[j], img)
                print(picture_name[i])

num = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24]
users = ['']#被实验者的名字
#        ,
for n in range(len(num)):
    for user in range(len(users)):
        print(users[user] + str(num[n]))
        img_path = 'E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/GAZE_DATA_backup/191202/'+users[user]+'/data_process/ROI/roi_picture_7_27/'+str(num[n])+'/'
        path_out = 'E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/GAZE_DATA_backup/191202/'+users[user]+'/data_process/ROI/clip_roi_picture_7_27/'+str(num[n])+'/'
        clip_image()

将得到的一秒60张图图片转为视频

import cv2
import os
from PIL import Image
from moviepy.editor import *
def videos():
    fps = 60
    fourcc = cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')
    video_writer = cv2.VideoWriter(filename= path_out + users[user]+ '_'+str(m+1) + '.mp4', fourcc=fourcc, fps=fps, frameSize=(100,100))
    for i in range(0,len(image_name)):
        img = imread(image_path + image_name[i], flags=cv2.IMREAD_COLOR, dtype=np.uint8)
        # cv2.imshow(str(image_name[i]),img)
        # img = cv2.imread(image_path + image_name[i])
        cv2.waitKey(100)
        # video_writer.write(img)
        print(str(image_name[i]) + ' done!' )
        video_writer.write(img)
      #   print(str(p) + '.png' + ' done!')
    video_writer.release()
num = [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24]
users = [''] #被实验者的名字

for user in range(len(users)):
    for n in range(len(num)):

        print(users[user] + str(num[n]))
        path_roi = 'E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/GAZE_DATA_backup/191202/'+users[user]+'/data_process/ROI/clip_roi_picture_7_27/'+str(num[n])+'/'
        for root, dirs, files in os.walk(path_roi):
            print(root)
            print(dirs)
            dirs = natsorted(dirs)
            for m in range(len(dirs)):
                image_path = root + str(dirs[m]) + '/'
                print(image_path)
                image_name = os.listdir(image_path)
                image_name = natsorted(image_name)
                print(image_name)
                path_out = 'E:/ffmpeg-latest-win64-static/2019_11_12_experimental_data/GAZE_DATA_backup/7_29/gaze/' + str(num[n]) + '_gaze/'
                mkdir(path_out)
                videos()