双目摄像头测距有部分代码没看懂

Question

在利用双目摄像头测距时，不仅可以测距还可以判断物体位于摄像头左边还是右边或者中间，下面是部分代码（用于判断左右的）没有看懂，还望各位可以解释一下
for *xyxy, conf, cls in reversed(det):
x = ((xyxy[2]-xyxy[0])/2)+xyxy[0]
y = ((xyxy[3]-xyxy[1])/2)+xyxy[1]
if(x<192):
pos='left'
elif((x<448) and (x>192)):
pos='middle'
else:
pos='right'
x=int(x.cpu())
y=int(y.cpu())

Answer 1

import argparse
import time
from pathlib import Path
import glob
import logging
import math
import os
import random
import shutil
import time
from itertools import repeat
from multiprocessing.pool import ThreadPool
from pathlib import Path
from threading import Thread
import cv2
import numpy as np
import torch
from PIL import Image, ExifTags
from torch.utils.data import Dataset
from tqdm import tqdm
from dis_count import *
from utils.general import xyxy2xywh, xywh2xyxy
from utils.torch_utils import torch_distributed_zero_first
import cv2
import torch
import torch.backends.cudnn as cudnn
from numpy import random

from models.experimental import attempt_load
from utils.datasets import LoadStreams, LoadImages
from utils.general import check_img_size, non_max_suppression, apply_classifier, scale_coords, xyxy2xywh,
strip_optimizer, set_logging, increment_path
from utils.plots import plot_one_box
from utils.torch_utils import select_device, load_classifier, time_synchronized
from dis_count import *
from utils.datasets import *

设置model

device = torch.device('cpu')
half = device.type != True # half precision only supported on CUDA

model = attempt_load('yolov5s.pt', map_location=device) # load FP32 model
imgsz = check_img_size(640, s=model.stride.max()) # check img_size

if half:
model.float() # to FP16

view_img = True
cudnn.benchmark = True # set True to speed up constant image size inference

Get names and colors

names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]

img01 = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img01.float() if half else img01) if device.type != 'cpu' else None # run once

cap1 = cv2.VideoCapture(0, cv2.CAP_DSHOW)
cap2 = cv2.VideoCapture(0, cv2.CAP_DSHOW)

while(True):

imgs = [None] * 1
imgs2 = [None] * 1

ref,a=cap1.read()
_,b=cap2.read()

# cv2.imshow('0', a)
# cv2.imshow('1', b)

cap1.grab()
cap2.grab()

_, imgs[0] = cap1.retrieve()
_, imgs2[0] = cap2.retrieve()

img = [letterbox(x, new_shape=640, auto=True)[0] for x in imgs]
# imgb = [letterbox(x1, new_shape=640, auto=True)[0] for x1 in imgs2]

# Stack
img = np.stack(img, 0)
# imgb = np.stack(imgb, 0)

# Convert
img = img[:, :, :, ::-1].transpose(0, 3, 1, 2)  # BGR to RGB, to bsx3x416x416
img = np.ascontiguousarray(img)

# imgb = imgb[:, :, :, ::-1].transpose(0, 3, 1, 2)  # BGR to RGB, to bsx3x416x416
# imgb = np.ascontiguousarray(imgb)


img = torch.from_numpy(img).to(device)
img = img.float() if half else img.float()  # uint8 to fp16/32
img /= 255.0  # 0 - 255 to 0.0 - 1.0   # torch.Size([1, 3, 480, 640])

# imgg = torch.from_numpy(imgb).to(device)
# imgg = imgg.half() if half else imgg.float()  # uint8 to fp16/32
# imgg /= 255.0  # 0 - 255 to 0.0 - 1.0

if img.ndimension() == 3:
    img = img.unsqueeze(0)
# if imgg.ndimension() == 3:
#     imgg = imgg.unsqueeze(0)

pred = model(img, augment=False)[0]
# predd = model(imgg, augment=False)[0]

# Apply NMS
pred = non_max_suppression(pred, 0.25, 0.45, classes=None, agnostic=False)
# predd = non_max_suppression(predd, 0.25, 0.45, classes=None, agnostic=False)

dislist,disp = dis_co(imgs[0], imgs2[0])
# dislist=torch.from_numpy(dislist)
def ved(pred):
    # t0 = time.time()
    for i, det in enumerate(pred):  # detections per image

        dis_box = dict()
        if True:  # batch_size >= 1
             s, im0 =  '%g: ' % i, imgs[i].copy()
        s += '%gx%g ' % img.shape[2:]  # print string
        gn = torch.tensor(im0.shape)[[1, 0, 1, 0]]  # normalization gain whwh
        if len(det):
            # Rescale boxes from img_size to im0 size
            det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()

            # Print results
            for c in det[:, -1].unique():
                n = (det[:, -1] == c).sum()  # detections per class
                s += '%g %ss, ' % (n, names[int(c)])  # add to string

            # Write results
            # print(det)
            dddd = 0
            d1=0
            for *xyxy, conf, cls in reversed(det):
                x = ((xyxy[2]-xyxy[0])/2)+xyxy[0]
                y = ((xyxy[3]-xyxy[1])/2)+xyxy[1]
                if(x<192):
                    pos='left'
                elif((x<448) and (x>192)):
                    pos='middle'
                else:
                    pos='right'
                x=int(x.cpu())
                y=int(y.cpu())
                # print(xyxy)
                dddd=(dislist[y][x]/5)[-1]
                label = '%s %.2f %.2f %s' % (names[int(cls)], conf, dddd,pos)
                msg={pos:dddd}
                dis_box.update(msg)
                # print(label)
                plot_one_box(xyxy, im0, label=label, color=colors[int(cls)], line_thickness=3)
    # print('Done. (%.3fs)' % (time.time() - t0))
    pos_msg=''
    dis_msg=0.
    for key, value in dis_box.items():
        if (value == min(dis_box.values())):
            print(key,value)
            # pos_msg=key
            # dis_msg=value
    return im0
def vedd(pred):
    for i, det in enumerate(pred):  # detections per image
        if True:  # batch_size >= 1
             s, im0 = '%g: ' % i, imgs2[i].copy()

        s += '%gx%g ' % img.shape[2:]  # print string
        gn = torch.tensor(im0.shape)[[1, 0, 1, 0]]  # normalization gain whwh
        if len(det):
            # Rescale boxes from img_size to im0 size
            det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()

            # Print results
            for c in det[:, -1].unique():
                n = (det[:, -1] == c).sum()  # detections per class
                s += '%g %ss, ' % (n, names[int(c)])  # add to string

            # Write results
            for *xyxy, conf, cls in reversed(det):
                label = '%s %.2f' % (names[int(cls)], conf)
                plot_one_box(xyxy, im0, label=label, color=colors[int(cls)], line_thickness=3)
    return im0

# 计算两针深度图  左帧img 右帧imgg
v1 = ved(pred)
# v2 = vedd(predd)
dis_box = dict()
dislist=np.ndarray(0)
cv2.imshow('0', v1)
# cv2.imshow('1', v2)
cv2.imshow('SGNM_disparity', (disp - 0) / 32)

c = cv2.waitKey(1) & 0xff
if c == 27:
    cap1.release()
    cap2.release()
    break