过程记录 yolov3目标检测输出目标坐标和原图物体像素坐标比较

设备：jetsonnano B01 一个usb广角摄像头
语言为python 库为opencv 框架pytorch
大家都知道yolov3的坐标变换过程比较复杂，大概要经过图片缩放，对数空间变换等等，然后在pytorch中是用一个tensor类型来表示张量的，所以一直认为目标检测的坐标是一个奇奇怪怪的值，需要经过变换才能得到原图目标的像素坐标
之前一直纠结怎么坐标转化，但今天实验之后怀疑其实检测结果的坐标转化为int数值以后跟实际原图像素坐标是一样的？
先上结果：
这是原始图片上的坐标点

左下角为像素坐标
（鼠标放在中间油渍的位置时显示的坐标

检测结果：

结果输出的坐标（转化为整数之后输出的）

两个坐标几乎一样，感觉其实它们两个是一个数值，而tensor类型后面多了一串小数点
完整代码
用opencv打开摄像头查看图片某点像素坐标：

import cv2
import os
import sys
import time
import datetime
import argparse
cap = cv2.VideoCapture(0)
i = 1while(cap.isOpened()):ret, frame = cap.read()cv2.imshow('frame',frame)if cv2.waitKey(30) == ord('q'):#ret, frame = cap.read()cv2.imwrite('data/custom/dd/'+str(i)+".jpg",frame) frame=cv2.imread('data/custom/dd/'+str(i)+".jpg")break

检测部分显示结果坐标：
pytorch python

        if detections is not None:# Rescale boxes to original imagedetections = rescale_boxes(detections, opt.img_size, img.shape[:2])unique_labels = detections[:, -1].cpu().unique()n_cls_preds = len(unique_labels)bbox_colors = random.sample(colors, n_cls_preds)for x1, y1, x2, y2, conf, cls_conf, cls_pred in detections:print("\t+ Label: %s, Conf: %.5f" % (classes[int(cls_pred)], cls_conf.item()))                              box_w = x2 - x1box_h = y2 - y1centx = int((x1+x2)*0.1/2)centy = int((y1+y2)*0.1/2)thex.append(centx)they.append(centy)str1+="x:"str1+=str(centx)str1+=",y:"str1+=str(centy)area = int(box_w*0.1)*int(box_h*0.1)if area > 600: time = 30print("x和y"+str((x2+x1)/2)+","+str((y1+y2)/2))print("x:")print(int((x1+x2)/2))print("y:")print(int((y1+y2)/2))print(int(box_w*0.1))print(int(box_h*0.1))color = bbox_colors[int(np.where(unique_labels == int(cls_pred))[0])]# Create a Rectangle patchbbox = patches.Rectangle((x1, y1), box_w, box_h, linewidth=2, edgecolor=color, facecolor="none")print(int(box_w)*int(box_h)*0.01)print(str1)if(box_w*box_h>50):se.write("1".encode())time.sleep(3)se.write("0".encode())se.write(str1.encode())data_read = se.readline()print(data_read.decode())# Add the bbox to the plotax.add_patch(bbox)# Add labelplt.text(x1,y1,s=classes[int(cls_pred)],color="white",verticalalignment="top",bbox={"color": color, "pad": 0},)# Save generated image with detectionsplt.axis("off")plt.gca().xaxis.set_major_locator(NullLocator())plt.gca().yaxis.set_major_locator(NullLocator())filename = os.path.basename(path).split(".")[0]#output_path = os.path.join("output", f"{filename}.png")#output_path = os.path.join("/data/custom/dd", f"{filename}.png")plt.savefig(filename, bbox_inches="tight", pad_inches=0.0)plt.close()

    if detections is not None:后面的代码是输出坐标的

关于目标检测的完整检测代码：（很长，熟悉的人不必看）

from __future__ import division
import serial as ser
import time
se=ser.Serial("/dev/ttyUSB0",115200,timeout=1)
from models import *
from utils.utils import *
from utils.datasets import *
from utils.augmentations import *
from utils.transforms import *
import cv2
import os
import sys
import time
import datetime
import argparsefrom PIL import Imageimport torch
import torchvision.transforms as transforms
from torch.utils.data import DataLoader
from torchvision import datasets
from torch.autograd import Variableimport matplotlib.pyplot as plt
import matplotlib.patches as patches
from matplotlib.ticker import NullLocator
i = 1
cap = cv2.VideoCapture(0)
array_of_img = []
start = time.time()
directory_name=r'output'while(cap.isOpened()):ret, frame = cap.read()cv2.imshow('frame',frame)if cv2.waitKey(30) == ord('q'):#ret, frame = cap.read()cv2.imwrite('data/custom/dd/'+str(i)+".jpg",frame) frame=cv2.imread('data/custom/dd/'+str(i)+".jpg")break
# When everything done, release the capturecap.release()cv2.destroyAllWindows()washtime = []
thex = []
they = []
area = 0
str1=""
if __name__ == "__main__":parser = argparse.ArgumentParser()parser.add_argument("--image_folder", type=str, default="data/samples", help="path to dataset")parser.add_argument("--model_def", type=str, default="config/yolov3.cfg", help="path to model definition file")parser.add_argument("--weights_path", type=str, default="weights/yolov3.weights", help="path to weights file")parser.add_argument("--class_path", type=str, default="data/coco.names", help="path to class label file")parser.add_argument("--conf_thres", type=float, default=0.8, help="object confidence threshold")parser.add_argument("--nms_thres", type=float, default=0.4, help="iou thresshold for non-maximum suppression")parser.add_argument("--batch_size", type=int, default=1, help="size of the batches")parser.add_argument("--n_cpu", type=int, default=0, help="number of cpu threads to use during batch generation")parser.add_argument("--img_size", type=int, default=416, help="size of each image dimension")parser.add_argument("--checkpoint_model", type=str, help="path to checkpoint model")opt = parser.parse_args()print(opt)device = torch.device("cuda" if torch.cuda.is_available() else "cpu")os.makedirs("output", exist_ok=True)# Set up modellamodel = Darknet(opt.model_def, img_size=opt.img_size).to(device)if opt.weights_path.endswith(".weights"):# Load darknet weightsmodel.load_darknet_weights(opt.weights_path)else:# Load checkpoint weightsmodel.load_state_dict(torch.load(opt.weights_path))model.eval()  # Set in evaluation modedataloader = DataLoader(ImageFolder(opt.image_folder, transform= \transforms.Compose([DEFAULT_TRANSFORMS, Resize(opt.img_size)])),batch_size=opt.batch_size,shuffle=False,num_workers=opt.n_cpu,)classes = load_classes(opt.class_path)  # Extracts class labels from fileTensor = torch.cuda.FloatTensor if torch.cuda.is_available() else torch.FloatTensorimgs = []  # Stores image pathsimg_detections = []  # Stores detections for each image indexprint("\nPerforming object detection:")prev_time = time.time()for batch_i, (img_paths, input_imgs) in enumerate(dataloader):# Configure inputinput_imgs = Variable(input_imgs.type(Tensor))# Get detectionswith torch.no_grad():detections = model(input_imgs)detections = non_max_suppression(detections, opt.conf_thres, opt.nms_thres)# Log progresscurrent_time = time.time()inference_time = datetime.timedelta(seconds=current_time - prev_time)prev_time = current_timeprint("\t+ Batch %d, Inference Time: %s" % (batch_i, inference_time))# Save image and detectionsimgs.extend(img_paths)img_detections.extend(detections)# Bounding-box colorscmap = plt.get_cmap("tab20b")colors = [cmap(i) for i in np.linspace(0, 1, 20)]print("\nSaving images:")# Iterate through images and save plot of detectionsfor img_i, (path, detections) in enumerate(zip(imgs, img_detections)):print("(%d) Image: '%s'" % (img_i, path))# Create plotimg = np.array(Image.open(path))plt.figure()fig, ax = plt.subplots(1)ax.imshow(img)# Draw bounding boxes and labels of detectionsif detections is not None:# Rescale boxes to original imagedetections = rescale_boxes(detections, opt.img_size, img.shape[:2])unique_labels = detections[:, -1].cpu().unique()n_cls_preds = len(unique_labels)bbox_colors = random.sample(colors, n_cls_preds)for x1, y1, x2, y2, conf, cls_conf, cls_pred in detections:print("\t+ Label: %s, Conf: %.5f" % (classes[int(cls_pred)], cls_conf.item()))                              box_w = x2 - x1box_h = y2 - y1centx = int((x1+x2)*0.1/2)centy = int((y1+y2)*0.1/2)thex.append(centx)they.append(centy)str1+="x:"str1+=str(centx)str1+=",y:"str1+=str(centy)area = int(box_w*0.1)*int(box_h*0.1)if area > 600: time = 30print("x和y"+str((x2+x1)/2)+","+str((y1+y2)/2))print("x:")print(int((x1+x2)/2))print("y:")print(int((y1+y2)/2))print(int(box_w*0.1))print(int(box_h*0.1))color = bbox_colors[int(np.where(unique_labels == int(cls_pred))[0])]# Create a Rectangle patchbbox = patches.Rectangle((x1, y1), box_w, box_h, linewidth=2, edgecolor=color, facecolor="none")print(int(box_w)*int(box_h)*0.01)print(str1)if(box_w*box_h>50):se.write("1".encode())time.sleep(3)se.write("0".encode())se.write(str1.encode())data_read = se.readline()print(data_read.decode())# Add the bbox to the plotax.add_patch(bbox)# Add labelplt.text(x1,y1,s=classes[int(cls_pred)],color="white",verticalalignment="top",bbox={"color": color, "pad": 0},)# Save generated image with detectionsplt.axis("off")plt.gca().xaxis.set_major_locator(NullLocator())plt.gca().yaxis.set_major_locator(NullLocator())filename = os.path.basename(path).split(".")[0]#output_path = os.path.join("output", f"{filename}.png")#output_path = os.path.join("/data/custom/dd", f"{filename}.png")plt.savefig(filename, bbox_inches="tight", pad_inches=0.0)plt.close()

串口：

se.write("50".encode())
print("缓存中的字节数")
print(se.inWaiting())
time.sleep(2)
se.write("100".encode())
print("缓存中的字节数")
print(se.inWaiting())#：返回接收缓存中的字节数
data = se.readline()
print(data.decode())
data = se.readline()
print(data.decode())
print(se.inWaiting())
se.write("my87u".encode())
print("缓存中的字节数")
print(se.inWaiting())
se.flush() #等待所有数据写出。
data = se.readline()
print(data.decode())
data = se.readline()
print(data.decode())

参数：

    parser = argparse.ArgumentParser()parser.add_argument("--image_folder", type=str, default="data/custom/dd2", help="path to dataset")parser.add_argument("--model_def", type=str, default="config/yolov3-custom.cfg", help="path to model definition file")parser.add_argument("--weights_path", type=str, default="checkpoints/ckpt_86.pth", help="path to weights file")parser.add_argument("--class_path", type=str, default="data/custom/classes.names", help="path to class label file")parser.add_argument("--conf_thres", type=float, default=0.8, help="object confidence threshold")parser.add_argument("--nms_thres", type=float, default=0.4, help="iou thresshold for non-maximum suppression")parser.add_argument("--batch_size", type=int, default=1, help="size of the batches")parser.add_argument("--n_cpu", type=int, default=0, help="number of cpu threads to use during batch generation")parser.add_argument("--img_size", type=int, default=416, help="size of each image dimension")parser.add_argument("--checkpoint_model",type=str,default="checkpoints/ckpt_86.pth", help="path to checkpoint model")opt = parser.parse_args()print(opt)

过程记录 yolov3目标检测输出目标坐标和原图物体像素坐标比较相关推荐

神经网络学习小记录50——Pytorch 利用efficientnet系列模型搭建yolov3目标检测平台
神经网络学习小记录50--Pytorch 利用efficientnet系列模型搭建yolov3目标检测平台学习前言什么是EfficientNet模型源码下载 EfficientNet模型的实现思 ...
【yolov3目标检测】(3) opencv+yolov3 检测交通路况，附python完整代码
各位同学好,今天和大家分享一下如何使用 opencv 调用 yolov3 模型,加载网络权重,很方便地实现 yolov3 目标检测.先放张图看效果. 使用的网上找的行车记录仪视频做测试,数据集采用CO ...
使用PyTorch从零开始实现YOLO-V3目标检测算法 (一)
原文:https://blog.csdn.net/u011520516/article/details/80222743 点击查看博客原文标检测是深度学习近期发展过程中受益最多的领域.随着技术的进步 ...
使用pytorch从零开始实现YOLO-V3目标检测算法 (二)
原文:https://blog.csdn.net/u011520516/article/details/80212960 博客翻译这是从零开始实现YOLO v3检测器的教程的第2部分.在上一节中,我 ...
使用PyTorch从零开始实现YOLO-V3目标检测算法 (三)
原文:https://blog.csdn.net/u011520516/article/details/80216009 点击查看博客原文这是从零开始实现YOLO v3检测器的教程的第3部分.第二部 ...
YOLOV3目标检测模型训练实例
YOLOV3目标检测从零开始学习使用keras-yolov3进行图片的目标检测,比较详细地记录了准备以及训练过程,提供一个信号灯的目标检测模型训练实例,并提供相关代码与训练集. DEMO测试 YOL ...
YOLOv3目标检测算法——通俗易懂的解析
目录 YOLOv3目标检测算法前沿一.YOLOv3 二.损失函数 YOLOv3目标检测算法前沿前两篇文章我们讲了下关于YOLOv1和YOLOv2的原理,有不懂的小伙伴可以回到前面再看看: YO ...
Pyqt搭建YOLOV3目标检测界面（超详细+源代码）
Pyqt搭建YOLOV3目标检测界面(超详细+源代码) 2022.5.25更新 2021.11.23 更新 2021.11.22 更新实现效果如下所示,可以检测图片.视频以及摄像头实时检测. 0.准 ...
深度篇——目标检测史(七) 细说 YOLO-V3目标检测之代码详解
返回主目录返回目标检测史目录上一章:深度篇--目标检测史(六) 细说 YOLO-V3目标检测下一章:深度篇--目标检测史(八) 细说 CornerNet-Lite 目标检测论文地址:< ...

过程记录 yolov3目标检测输出目标坐标和原图物体像素坐标比较

过程记录 yolov3目标检测输出目标坐标和原图物体像素坐标比较相关推荐

最新文章

热门文章