105 lines
4.0 KiB
Python
105 lines
4.0 KiB
Python
from icraft.xir import *
|
|
from icraft.xrt import *
|
|
from icraft.host_backend import *
|
|
from icraft.buyibackend import *
|
|
import numpy as np
|
|
import pandas as pd
|
|
from numpy.linalg import norm
|
|
from typing import List
|
|
import platform
|
|
import cv2
|
|
import json
|
|
import os
|
|
from tqdm import tqdm
|
|
from kornia import tensor_to_image, image_to_tensor
|
|
from kornia.color import ycbcr_to_rgb, rgb_to_bgr, rgb_to_ycbcr, bgr_to_rgb
|
|
import torch
|
|
from torchvision.transforms import Resize
|
|
|
|
|
|
def sim(network :Network, input: List[Tensor]) -> List[Tensor]:
|
|
session = Session.Create([ HostBackend ], network.view(0), [ HostDevice.Default() ])
|
|
session.enableTimeProfile(True)
|
|
session.apply()
|
|
output_tensors = session.forward( input )
|
|
return output_tensors
|
|
|
|
|
|
def run(network: Network, input: List[Tensor]) -> List[Tensor]:
|
|
URL_PATH=""
|
|
if platform.machine() == 'aarch64':
|
|
URL_PATH = R"axi://ql100aiu?npu=0x40000000&dma=0x80000000";
|
|
else:
|
|
URL_PATH = Rf"socket://ql100aiu@192.168.125.{ip_octet}:9981?npu=0x40000000&dma=0x80000000"
|
|
# Buyi_device = Device.Open(URL_PATH)
|
|
session = Session.Create([ HostBackend], network.view(0), [HostDevice.Default()])
|
|
# session = Session.Create([ BuyiBackend, HostBackend], network.view(0), [ Buyi_device,HostDevice.Default()])
|
|
session.enableTimeProfile(True) #打开计时功能
|
|
session.apply()
|
|
output_tensors = session.forward( input ) #前向
|
|
|
|
# 计算时间
|
|
result = session.timeProfileResults() #获取时间，[总时间，传输时间, 硬件时间，余下时间]
|
|
# save_time(network, result, result_table_name)
|
|
time = np.array(list(result.values()))
|
|
total_softtime = np.sum(time[:,1])
|
|
total_hardtime = np.sum(time[:,2])
|
|
# print("total_softtime:{},total_hardtime:{}(ms)",total_softtime,total_hardtime)
|
|
return output_tensors
|
|
|
|
|
|
def gray_read(img_path):
|
|
img_n = cv2.imread(str(img_path), cv2.IMREAD_GRAYSCALE)
|
|
img_t = image_to_tensor(img_n).float() / 255
|
|
return img_t
|
|
|
|
|
|
def ycbcr_read(img_path):
|
|
img_n = cv2.imread(str(img_path), cv2.IMREAD_COLOR)
|
|
img_t = image_to_tensor(img_n).float() / 255
|
|
img_t = rgb_to_ycbcr(bgr_to_rgb(img_t))
|
|
y, cbcr = torch.split(img_t, [1, 2], dim=0)
|
|
return y, cbcr
|
|
|
|
ip_octet = 93
|
|
GENERATED_JSON_FILE = "../2_compile/imodel/tardal_dt/tardal_dt_BY.json"
|
|
GENERATED_RAW_FILE = "../2_compile//imodel/tardal_dt/tardal_dt_BY.raw"
|
|
|
|
generated_network = Network.CreateFromJsonFile(GENERATED_JSON_FILE)
|
|
generated_network.loadParamsFromFile(GENERATED_RAW_FILE)
|
|
|
|
|
|
vi_path = '../0_TarDAL-main/data/m3fd/vi/00000.png'
|
|
ir_path = '../0_TarDAL-main/data/m3fd/ir/00000.png'
|
|
ori_img = cv2.imread(vi_path)
|
|
vi, cbcr = ycbcr_read(vi_path)
|
|
ir = gray_read(ir_path)
|
|
max_size = torch.Size([768, 1024])
|
|
transform_fn = Resize(size=max_size)
|
|
t = torch.cat([ir, vi, cbcr], dim=0)
|
|
ir, vi, cbcr = torch.split(transform_fn(t), [1, 1, 2], dim=0)
|
|
ir = ir.unsqueeze(-1).numpy().transpose(3, 1, 2, 0)
|
|
vi = vi.unsqueeze(-1).numpy().transpose(3, 1, 2, 0)
|
|
cbcr = cbcr.unsqueeze(-1).numpy().transpose(3, 1, 2, 0)
|
|
# ir.tofile('./ir.bin')
|
|
# vi.tofile('./vi.bin')
|
|
# cbcr.tofile('./cbcr.bin')
|
|
|
|
# ir = np.fromfile('./ir.bin', dtype=np.float32).reshape((1, 768, 1024, 1))
|
|
# vi = np.fromfile('./vi.bin', dtype=np.float32).reshape((1, 768, 1024, 1))
|
|
# cbcr = np.fromfile('./cbcr.bin', dtype=np.float32).reshape((1, 768, 1024, 2))
|
|
input1, input2 = Tensor(ir, Layout("NHWC")), Tensor(vi, Layout("NHWC"))
|
|
generated_output = run(generated_network, [input1, input2])
|
|
output = generated_output[0]
|
|
|
|
fuse = np.concatenate((output, cbcr), axis=3).transpose(0, 3, 1, 2)
|
|
# fuse.tofile('./fuse.bin')
|
|
# fuse = np.fromfile('./fuse.bin', dtype=np.float32).reshape((1, 768, 1024, 3)).transpose(0, 3, 1, 2)
|
|
img = ycbcr_to_rgb(torch.from_numpy(fuse))
|
|
img = rgb_to_bgr(img)
|
|
img_n = tensor_to_image(img.squeeze().cpu()) * 255
|
|
if ori_img.shape != img_n.shape:
|
|
img_n = cv2.resize(img_n, (ori_img.shape[1], ori_img.shape[0]), interpolation=cv2.INTER_LINEAR)
|
|
# cv2.imwrite('./fuse.png', img_n)
|
|
cv2.imshow(" ", img_n)
|
|
cv2.waitKey(0) |