<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[Daweixi的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/Daweixi</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; Daweixi]]></copyright><item><title><![CDATA[error: Could not find suitable distribution for Requirement.parse(‘tf-slim＞=1.1‘)]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/108790307</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/108790307</guid><author>Daweixi</author><pubDate>Fri, 25 Sep 2020 10:07:36 +0800</pubDate><description><![CDATA[笔者在部署windows 版本tensorflow object detection API的时候，参照如下博客：

https://blog.youkuaiyun.com/qq_40434140/article/details/88083142

这篇博客的步骤已经十分得求稳了，但在最后一步还是出现了如标题所示得问题。



解决方法其实已经在报错中有提示了。

看上面第三行，有tf-slim的地址：https://pypi.python.org/simple/tf-slim/

打开之后只有两个链接，英文再不好的人]]></description><category></category></item><item><title><![CDATA[判断文件或文件夹是否存在的shell脚本代码]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/108365344</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/108365344</guid><author>Daweixi</author><pubDate>Wed, 02 Sep 2020 16:46:34 +0800</pubDate><description><![CDATA[转载自http://www.cppcns.com/os/linux/92476.html

判断文件是否存在的shell脚本，有详细的注释，很不错，有需要的朋友不妨参考下

实现代码一、


#!/bin/sh
# 判断文件是否存在
# link：www.jb51.net
# date：2013/2/28
 
myPath="/var/log/httpd/"
myFile="/var /log/httpd/access.log"
 
# 这里的-x 参数判断$myPath是否存在并且是否具有可执行权限
if]]></description><category></category></item><item><title><![CDATA[opencv保存mp4文件格式，出现问题提示OpenCV: FFMPEG: fallback to use tag 0x7634706d/‘mp4v‘]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/107958572</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/107958572</guid><author>Daweixi</author><pubDate>Wed, 12 Aug 2020 15:31:32 +0800</pubDate><description><![CDATA[https://blog.youkuaiyun.com/qq_39877362/article/details/105179245?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522159721680319195264534996%2522%252C%2522scm%2522%253A%252220140713.130102334.pc%255Fall.%2522%257D&amp;request_id=159721680319195264534996&am]]></description><category></category></item><item><title><![CDATA[opencv+python合并视频，可选择起始帧]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/107768969</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/107768969</guid><author>Daweixi</author><pubDate>Mon, 03 Aug 2020 17:14:47 +0800</pubDate><description><![CDATA[import cv2

fourcc = cv2.VideoWriter_fourcc(*'XVID')
out = cv2.VideoWriter('merged_drum.avi',fourcc, 25.0, (1920, 1080))

for path in ['new1.avi','new4.avi','new2.avi']: # 需要合并的视频名称，也可以用os.listdir() 
    cap = cv2.VideoCapture(path)
    # frameToStart = 1.]]></description><category></category></item><item><title><![CDATA[/usr/bin/ld: warning: libhdf5_hl.so.100, needed by .build_release/lib/libcaffe.so, not found (try us]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/107165956</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/107165956</guid><author>Daweixi</author><pubDate>Mon, 06 Jul 2020 19:24:12 +0800</pubDate><description><![CDATA[sudo cp -s $HOME/anaconda3/lib/libhdf5_hl.so.100.1.0 /usr/lib/libhdf5_hl.so.100
sudo cp -s $HOME/anaconda3/lib/libhdf5_hl.so.100.1.0 /usr/lib/x86_64-linux-gnu/libhdf5_hl.so.100
sudo cp -s $HOME/anaconda3/lib/libhdf5.so.101.1.0 /usr/lib/libhdf5.so.101
sudo.]]></description><category></category></item><item><title><![CDATA[linux下安装leveldb（非github源代码方法），解决/usr/bin/ld: 找不到 leveldb]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/107160021</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/107160021</guid><author>Daweixi</author><pubDate>Mon, 06 Jul 2020 15:29:42 +0800</pubDate><description><![CDATA[首先给出我最终安装成功的参考链接

https://ywnz.com/linuxysjk/4280.html

其次附上参考过的其他方式链接，大同小异。基本是从github上下载源代码，然后cmake，再make。但是我这边出现一个问题就是cmake一直不能成功。

https://blog.youkuaiyun.com/a1165741556/article/details/104028855

https://www.cnblogs.com/xueqiuqiu/p/8268814.html

https://blo]]></description><category></category></item><item><title><![CDATA[为什么L1正则化会有稀疏性？为什么L1正则化能进行内置特征选择？]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/106526107</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/106526107</guid><author>Daweixi</author><pubDate>Wed, 03 Jun 2020 16:09:38 +0800</pubDate><description><![CDATA[# 首先给大家推荐一个比较直观地搞懂L1和L2正则化的思考，有视频有图像，手动赞！

https://zhuanlan.zhihu.com/p/25707761

当然也推荐看看https://www.cnblogs.com/jclian91/p/9824310.html

回到问题，为什么L1正则化会有稀疏性？先盗个图。



看右边那个图，可以发现L1正则化的黄色部分，其第一象限的斜率是固定的。要使得J(θ)在坐标轴之外的地方和L1正则化的框框相交，只有在J(θ)的长轴的斜率是-1的时候(为了方便说明，]]></description><category></category></item><item><title><![CDATA[linux系统安装firefox的flash player插件]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/106175133</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/106175133</guid><author>Daweixi</author><pubDate>Sun, 17 May 2020 15:36:53 +0800</pubDate><description><![CDATA[1.下载插件并解压插件

http://get.adobe.com/cn/flashplayer/

我下载的是：install_flash_player_11_linux.i386.tar.gz

在下载目录打开终端，输入命令："tar -zxvf FlashPlayer.tar.gz"，

解压完成后会得到："libflashplayer.so"文件，这就是Flash插件。

2.把上述插件复制到正确目录

在终端继续输入下面两行代码。第一行是复制插件到火狐的插件目录，第二行是给权限


sudo cp]]></description><category></category></item><item><title><![CDATA[linux系统用conda安装opencv，填坑与方法汇总，解决Solving environment: failed with initial frozen solve等问题]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/106148290</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/106148290</guid><author>Daweixi</author><pubDate>Sat, 16 May 2020 11:50:59 +0800</pubDate><description><![CDATA[conda install -c https://conda.anaconda.org/menpo opencv3 
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from curren.]]></description><category></category></item><item><title><![CDATA[python中出现invalid syntax报错]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/106048579</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/106048579</guid><author>Daweixi</author><pubDate>Mon, 11 May 2020 10:23:04 +0800</pubDate><description><![CDATA[class Solution:
    def myPow(self, x: float, n: int) -&gt; float:
        if x==0:
            return 0
        if n&gt;0:
            return x*(self.myPow(x,n-1))
        if n==0:
            return 1
        if n&lt;0:
            return self.myPow(1/x.]]></description><category></category></item><item><title><![CDATA[ffmpeg：RuntimeError: No MovieWriters available!]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/104896961</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/104896961</guid><author>Daweixi</author><pubDate>Mon, 16 Mar 2020 13:03:55 +0800</pubDate><description><![CDATA[参考了以下两篇博文：

https://blog.youkuaiyun.com/weixin_39278265/article/details/84147458

https://blog.youkuaiyun.com/mys_mys/article/details/99646476

具体解决：

第一步：从下面链接下载来ffmpeg，然后解压

http://ffmpeg.zeranoe.com/builds/

h...]]></description><category></category></item><item><title><![CDATA[opencv+python（c++）多轮廓质心，孤立区域质心]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/104443507</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/104443507</guid><author>Daweixi</author><pubDate>Sat, 22 Feb 2020 14:37:30 +0800</pubDate><description><![CDATA[# 求各孤立区域质心
sure_fg = np.array(sure_fg,np.uint8)
fg_b,contours,hierarch=cv2.findContours(sure_fg,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_NONE)# sure_fg要灰度图

mu = []
mc = []
for i in range(len(contours)):
 ...]]></description><category></category></item><item><title><![CDATA[Visual Studio打开文件时出现“向程序发送命令时出现问题‘’]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/103712501</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/103712501</guid><author>Daweixi</author><pubDate>Thu, 26 Dec 2019 11:33:11 +0800</pubDate><description><![CDATA[这可能是之前用其他格式打开过文件，比如记事本或者notepad啥的。

关掉打开的错误文件，然后从打开方式那边选择用vs打开试试。

我是这样处理了一下，就可以了
...]]></description><category></category></item><item><title><![CDATA[分水岭算法原理和python+opencv实现，十分详细]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/103570568</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/103570568</guid><author>Daweixi</author><pubDate>Mon, 16 Dec 2019 21:28:00 +0800</pubDate><description><![CDATA[https://www.cnblogs.com/ssyfj/p/9278815.html

结果图中有个红色的框，不知道咋解决
]]></description><category></category></item><item><title><![CDATA[Image.getdata(band=None)]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/102600532</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/102600532</guid><author>Daweixi</author><pubDate>Thu, 17 Oct 2019 10:23:01 +0800</pubDate><description><![CDATA[Image.getdata(band=None)

函数的作用是把图片的像素信息flatten到一维，搞成一个特征向量的形式，当然这边并没有进行特征提取，只是直接把所有像素做成了向量。

参数band可以理解为通道，如果输入为0的话就是会输出R 通道像素组成的像素信息。
...]]></description><category></category></item><item><title><![CDATA[FindCUDA.cmake位置；OpenCVDetectCUDA.cmake位置；cuda_fp16.h位置]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/102556449</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/102556449</guid><author>Daweixi</author><pubDate>Mon, 14 Oct 2019 20:50:18 +0800</pubDate><description><![CDATA[FindCUDA.cmake位置：E:\opencv3.1\tool\cmake-3.8.0-win64-x64\share\cmake-3.8\Modules

OpenCVDetectCUDA.cmake位置：E:\opencv3.1\opencv\sources\cmake

cuda_fp16.h位置：C:\Program Files\NVIDIA GPU Computing Toolki...]]></description><category></category></item><item><title><![CDATA[np.transpose和np.swapaxes关于维度的一些思考]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/100981400</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/100981400</guid><author>Daweixi</author><pubDate>Wed, 18 Sep 2019 14:08:47 +0800</pubDate><description><![CDATA[先把我们的矩阵输出，有一个比较直观的印象。接下来理解0，1，2.

实际上，一般的入门选手应该已经习惯从0开始数数了，这边也应该继续有这个思维习惯。0是我们可以暂时忽略的一层，因为是深度层。

所以我们在看transpose之后的形状，只需要从下面这个二维矩阵看就好了，向下为1，向右为2.



transpose成(2,1,0)后，就是从三维空间中往右看。最终得到如下：




...]]></description><category></category></item><item><title><![CDATA[python中yield详解，适合入门]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/100656206</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/100656206</guid><author>Daweixi</author><pubDate>Mon, 09 Sep 2019 11:21:08 +0800</pubDate><description><![CDATA[之前完全没有看过yield，也不知道啥生成器，看过这篇基本能懂了。plus：很多其他的帖子，科科

python2：

https://www.runoob.com/w3cnote/python-yield-used-analysis.html

Python3迭代器和生成器：

https://www.runoob.com/python3/python3-iterator-generator.h...]]></description><category></category></item><item><title><![CDATA[RGB转换YUV，lab，hsv等通道，并绘制直方图]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/100566704</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/100566704</guid><author>Daweixi</author><pubDate>Thu, 05 Sep 2019 20:40:03 +0800</pubDate><description><![CDATA[import cv2
import numpy as np
import matplotlib.pyplot as plt
img = cv2.imread("../../picture/17.jpg",1)
# print(img.shape)
img_hsv = cv2.cvtColor(img,cv2.COLOR_RGB2HSV)
img_Lab = cv2.cvtColor(img,cv...]]></description><category></category></item><item><title><![CDATA[python灰度共生矩阵GLCM，带公式和代码]]></title><link>https://blog.youkuaiyun.com/Daweixi/article/details/100553172</link><guid>https://blog.youkuaiyun.com/Daweixi/article/details/100553172</guid><author>Daweixi</author><pubDate>Thu, 05 Sep 2019 10:45:47 +0800</pubDate><description><![CDATA[灰度共生矩阵所代表的含义是图像的灰度值在空间的某个方向上的特点，但是如何衡量两个目标图像纹理之间的差异是不能直接使用灰度共生矩阵表示出来的，所以需要使用一个特定的模型去统计灰度共生矩阵的特性来描述纹理特性。例如Haralick HaralickHaralick曾提出了14种基于灰度共生矩阵计算出来的统计量：能量、熵、对比度、均匀性、相关性、方差、和平均、和方差、和熵、差方差、差平均、差熵、...]]></description><category></category></item></channel></rss>