highoooo的博客

图像开运算是图像依次经过腐蚀、膨胀处理后的过程。图像被腐蚀后，去除了噪声，但是也压缩了图像；接着对腐蚀过的图像进行膨胀处理，可以去除噪声，并保留原有图像。开运算实际是先腐蚀运算，再膨胀运算，把看上去把细微连在一起的两块目标分开了。一般来说，开运算可以使图像的轮廓变得光滑，还能使狭窄的连接断开和消除细毛刺。开运算在过滤噪声的同时并没有对物体的形状、轮廓造成明显的影响，这是一大优势。当只关心物体的位置或者个数时，物体形状的改变不会给任务带来困扰，此时用开运算处理具有处理速度上的优势。闭运算是开运算的相反操作，先

sort

https://docs.python.org/zh-cn/3.7/library/bisect.html

https://blog.csdn.net/qq_40824311/article/details/103017760

from torch import nnimport torchtorch.manual_seed(21)input = torch.randn(1,3,3,3).cuda()input[0][0] = 0m3 = nn.BatchNorm2d(3, eps=0, momentum=0.5, affine=True, track_running_stats=True).cuda()m3.running_mean = (torch.ones([3])*4).cuda() # 设置模型的均

from torchvision.models import resnet50from thop import profileimport torchmodel = resnet50()input = torch.randn(1, 3, 224, 224)flops, params = profile(model, inputs=(input, ))print("%s ------- params: %.2fMB ------- flops: %.2fG" % (model, params /

侵线检测class cal_tools(object): def cal_slope(self,line): coorA = (line[0][0], line[0][1]) coorB = (line[0][2], line[0][3]) m = (coorB[1] - coorA[1]) / (coorB[0] - coorA[0] + 0.0000001) return m def slope_filter(self,

import cv2import osimport numpy as npimport time# calculate The Phase Only Transformdef PHOT(image): # DFT F = np.fft.fft2(image) # get magnitude M = np.abs(F) # Get phase P = F / M # Reconstruct image from phase R =

import scipy.statsimport torchimport globimport cv2import randomimport numpy as npdef make_noise(image_size, diag_limit=10): a = np.pi * np.random.random() R = np.array([ [np.cos(a), np.sin(a)], [-np.sin(a), np.cos(a)],

一、pytorch模型转为tensorrt模型1.yololayer.h中修改class_num2.Cmakelist中修改cuda路径3.build中编译生成yolov5文件4.sudo ./yolov5 -s [.wts] [.engine] [s/m/l/x/s6/m6/l6/x6 or c/c6 gd gw] // serialize model to plan file 生成.engine文件5.sudo ./yolov5 -d yolov5.engine …/samples 确定

vars() : 与args和class相关 转为字典形式星号 https://blog.csdn.net/zkk9527/article/details/88675129a={'A':1,'B':2,'C':3}print(a)print(*a)-------------------{'A': 1, 'B': 2, 'C': 3}A B CPath：https://blog.csdn.net/hxj0323/article/details/113374539...

效果演示：（顺时针方向 - Ori-image/After canny/Final area/Hough lines）Maindef getline(imgpath): save_Edges = True save_Lines = True # Default - rate = 0.25 img, general_point = img_Preprocess(imgpath) gray, edges, orgb, orgb_lines, lines =

import cv2import numpy as npimport collections# 定义字典存放颜色分量上下限# 例如：{颜色: [min分量, max分量]}# {'red': [array([160, 43, 46]), array([179, 255, 255])]}def getColorList(): dict = collections.defaultdict(list) # 黑色 lower_black = np.array([0,

一、删除休眠文件：转到“开始”菜单，键入“cmd”，然后右键单击结果“命令提示符”，然后选择“以管理员身份运行”。然后键入“powercfg. exe -h off”，然后按Enter。之后，你应该在C盘上看到多几个GB的可用空间。二、删除临时文件：Appdata/Local/temp 里面文件删除三、删除Local不必要文件：Appdata/Local 里面不必要文件删除四、开始->系统清理C盘文件五、C盘->属性->清理系统文件....

https://blog.csdn.net/guyuealian/article/details/88343924

https://www.pynote.net/archives/1621#:~:text=%E8%80%8Cadd_argument%E5%87%BD%E6%95%B0%E6%9C%89%E4%B8%80%E4%B8%AAnargs%E5%8F%82%E6%95%B0%EF%BC%8C%E9%80%9A%E8%BF%87%E6%AD%A4%E5%8F%82%E6%95%B0%EF%BC%8C%E5%8F%AF%E4%BB%A5%E5%AE%9E%E7%8E%B0%E5%91%BD%E4%BB%A4%E8%A

import matplotlib.pyplot as pltimport globimport osimport numpy as npimport pylab as plfile_names = os.listdir("824_4classes_200epoch_logs")train_Loss_merge = np.empty(len(file_names))val_Loss_merge = np.empty(len(file_names))Epoch_merge = np.a

xml -> annotationclass提前设定好import xml.etree.ElementTree as ETimport randomimport globall_class = ['straw hat' , 'train' , 'person' , 'hat' , 'head' , 'cellphone' , 'cell phone']act_class = ['hat' , 'head' , 'cellphone']def Read_Write_classest.

class Data_test(object): day=0 month=0 year=0 def __init__(self,year=0,month=0,day=0): self.day=day self.month=month self.year=year def out_date(self): print "year :" print self.year prin

import cv2import numpy as npimport requestsimport urllib.requestimport requestsimport jsondef my_post(): data = {"imageID":(None,"1648"), "imageKeyWord":(None,"811"), "project":(None,"bolttest"), "imgFileA":("

cv2.boundingRect和cv2.fillPol 用多边形填充图形： pts：提取出来的特征点 转为np.array->astype(dtype=int)->reshape(-1,1,2) ///boundingRec返回正矩形 /// minAreaRect返回最小斜矩形 x,y,w,h：返回矩形框的x,y,w,h newImg = np.full(img.shape, 255, dtype=np.uint8) 创建一个背景图

1.files = os.listdir(……)顺序读取：files = [‘img1.jpg’,‘img5.jpg’,‘img9.jpg’]for i in range(len(files)):list = [f"img{i}".jpg ]return listos.pat.exists(file)+continue若不存在文件则进入下一个循环当图片是img0.jpg ,img5.jpg可用...

import randomwith open('annotation.txt', 'r') as ff: dataList = ff.readlines() tempList = set(range(len(dataList)))valIndex = set(random.sample(tempList, 760))trainIndex = tempList - valIndexvalList = [dataList[i].strip() for i in list(valInd

https://blog.csdn.net/qq_38883271/article/details/96439208lenhttps://www.zhihu.com/question/46973549/answer/767530541inithttps://blog.csdn.net/chituozha5528/article/details/78354833getitem

https://blog.csdn.net/Haiqiang1995/article/details/89284201

https://blog.csdn.net/weixin_36279318/article/details/79176475