def readVideo(RGB_videoname, th_videoname, outputFolder):
time_ori = getOriginaltime(RGB_videoname, th_videoname)
# thermal processing
frame_IR = imageio.get_reader(th_videoname, 'ffmpeg')
IR_nframes = frame_IR._meta['nframes']
# RGB processing
frame_RGB = imageio.get_reader(RGB_videoname, 'ffmpeg')
RGB_nframes = frame_RGB._meta['nframes']
for th_num in range(1195, 1197):
sec_add = th_num / frame_IR._meta['fps']
#sec_add = 13.265
RGB_num = int((float(th_num) / IR_nframes) * RGB_nframes)
image_IR = frame_IR.get_data(th_num) #480*640*3
image_RGB = frame_RGB.get_data(RGB_num) #1080*1920*3
# plt.imshow(image_RGB)
# plt.show()
img_temp = image_RGB.copy()
bounding_boxes_th, bounding_boxes_RGB, human = Detection.Integrate(
image_IR, image_RGB)
bounding_boxes_B, object = Detection.hsv_thresholding(image_RGB)
time_final = time_ori + datetime.timedelta(
hours=0, minutes=0, seconds=sec_add)
time_final_string = getTimestr(time_final)
if human is True:
########################################################
#############################thermal#####################################
pixel_T = np.zeros((bounding_boxes_th.shape[0], 2)) #n*2
for i in range(bounding_boxes_th.shape[0]):
# img_temp = cv2.rectangle(img_temp, (bounding_boxes_th[i][0], bounding_boxes_th[i][1]),
# (bounding_boxes_th[i][0] + bounding_boxes_th[i][2],
# bounding_boxes_th[i][1] + bounding_boxes_th[i][3]),
# (255, 0, 0), 1)
# plt.imshow(img_temp)
# plt.show()
pixel_T[i, :] = np.array(
(bounding_boxes_th[i, 2] / 2 + bounding_boxes_th[i, 0],
bounding_boxes_th[i, 3] + bounding_boxes_th[i, 1]))
text_file_th.write(
"%s %s %s %s" %
(time_final_string, str(pixel_T[i, 0]), str(
pixel_T[i, 1]), "H\n")) #thermal folder + txt'H'
#plt.savefig()
########################################################
#################################RGB#####################################
pixel_R = np.zeros((bounding_boxes_RGB.shape[0], 2)) # n*2
for i in range(bounding_boxes_RGB.shape[0]):
bounding_boxes_RGB[i][0] += 106
bounding_boxes_RGB *= 1080 / 480
pixel_R[i, :] = np.array(
(bounding_boxes_RGB[i, 2] / 2 + bounding_boxes_RGB[i, 0],
bounding_boxes_RGB[i, 3] + bounding_boxes_RGB[i, 1]))
text_file_RGB.write(
"%s %s %s %s" %
(time_final_string, str(pixel_R[i, 0]), str(
pixel_R[i, 1]), "H\n")) #RGB folder + txt'H'
# img_temp = cv2.rectangle(img_temp, (bounding_boxes_RGB[i][0], bounding_boxes_RGB[i][1]),
# (bounding_boxes_RGB[i][0] + bounding_boxes_RGB[i][2],
# bounding_boxes_RGB[i][1] + bounding_boxes_RGB[i][3]),
# (255, 0, 0), 1)
# plt.imshow(img_temp)
# plt.show()
#plt.savefig()
if object is True:
pixel_R_B = np.zeros((bounding_boxes_B.shape[0], 2)) # n*2
for i in range(bounding_boxes_B.shape[0]):
pixel_R_B[i, :] = np.array(
(bounding_boxes_B[i, 2] / 2 + bounding_boxes_B[i, 0],
bounding_boxes_B[i, 3] / 2 + bounding_boxes_B[i, 1]))
text_file_th.write(
"%s %s %s %s" %
(time_final_string, str(pixel_R_B[i, 0]),
str(pixel_R_B[i, 1]), "H\n")) #RGB folder + txt'B'
