Kezhi's Blog

Vue

Notes on Using Vue with Flask

There may be conflicts when using Vue with Flask, but they can be resolved by changing the configuration.

• Flask

1
2
3
4

if __name__ == '__main__':
    app.jinja_env.variable_start_string = '[['
    app.jinja_env.variable_end_string = ']]'
    app.run(debug=True)

• Vue

1
2
3

new Vue({
    delimiters: ['[[',']]']
})

Directives

请求伪装

在爬取腾讯天气时，发现无法成功爬取到天气数据，经过尝试发现是请求头除了问题。爬虫在爬取网页时的默认请求头为：

这种默认的请求头在爬取很多网站时会被禁止，无法获取到网页数据，因此需要使用请求头伪装，让网站认为是浏览器在访问。操作步骤如下：

Request Spoofing

When crawling Tencent weather data, it was found that the weather data could not be successfully crawled. After some attempts, it was discovered that the issue lies in the request headers. The default request headers used by the spider when crawling web pages are:

This default request header may be blocked when crawling many websites, making it impossible to obtain web page data. Therefore, request header spoofing is needed to make the website think it is a browser accessing it. The operation steps are as follows:

简介

Bilibili（以下简称B站）中有大量的番剧版权，截止目前一共有3161部。每一部番剧都可以找到它的播放量，追番量，弹幕数量等播放数据，除此之外，每部番剧还有其相应的标签（如“漫画改”，“热血”，“搞笑”）。本项目旨在分析番剧播放数据与番剧标签之间的关系，同时也是一项数据分析的大作业，采用APriori频繁项集挖掘进行分析。

GitHub地址：https://github.com/KezhiAdore/BilibiliAnimeData_Analysis

码云地址：https://gitee.com/KezhiAdore/BilibiliAnimeData_Analysis

Introduction

Bilibili (referred to as B station) has a large number of anime copyrights, with a total of 3161 as of now. Each anime can be found with its play count, follow count, barrage count, and other playback data. In addition, each anime has its corresponding tags (such as "comic adaptation", "hot-blooded", "comedy"). This project aims to analyze the relationship between anime playback data and anime tags, and it is also a data analysis project that uses APriori frequent itemset mining for analysis.

GitHub address: https://github.com/KezhiAdore/BilibiliAnimeData_Analysis

Gitee address: https://gitee.com/KezhiAdore/BilibiliAnimeData_Analysis

GitHub地址：https://github.com/KezhiAdore/MultiRobots_CoverMap

码云地址：https://gitee.com/KezhiAdore/MultiRobots_CoverMap

一、问题提出

1. 背景

机器人在工业、国防和科学技术中的应用日益广泛， 它能够在枯燥和危险的复杂非结构化环境下工作。使用机器人大大提高了人们的工作效率， 改变了人们的生活方式， 带来了巨大的经济和社会效益， 有力地推动了有关学科和技术领域的发展。

区域覆盖是指携带有一定探测范围的传感器如激光、声纳等的机器人探索访问整个区域， 并完成相应任务的过程。区域覆盖是自主移动机器人的一项基本任务。

GitHub link: https://github.com/KezhiAdore/MultiRobots_CoverMap

Gitee link: https://gitee.com/KezhiAdore/MultiRobots_CoverMap

1. Introduction

1. Background

Robots are increasingly being used in industries, defense, and scientific and technological fields. They can work in complex and dangerous environments that are monotonous and unstructured. The use of robots greatly improves work efficiency, changes people's lifestyles, brings enormous economic and social benefits, and promotes the development of related disciplines and technologies.

Area coverage refers to the process of exploring and accessing the entire area with robots carrying sensors with a certain detection range, such as lasers and sonars, and completing corresponding tasks. Area coverage is a fundamental task of autonomous mobile robots.

频域滤波器简介

频域滤波器与空域滤波器相对，空域滤波器是空间卷积块与空间图像做卷积运算，而频域滤波是在将图像进行了离散傅里叶变换之后在频域上与滤波器做乘法，之后再进行傅里叶反变换得到滤波之后的图像。

使用频域滤波器对图像进行滤波的基本步骤：

• 对图像进行DFT变换得到原图像的频域图像
• 频域图像与频域滤波器相乘得到新的频域图像
• 对新的频域图像进行IDFT变换到空域得到新图像。

Introduction to Frequency Domain Filters

Frequency domain filters are different from spatial domain filters. Spatial domain filters perform convolution operations between spatial convolution kernels and spatial images, while frequency domain filters perform multiplication operations between the frequency domain image obtained by discrete Fourier transform (DFT) and the filter in the frequency domain, and then perform inverse Fourier transform (IDFT) to obtain the filtered image.

The basic steps of using frequency domain filters to filter images are as follows:

• Perform DFT on the image to obtain the frequency domain image of the original image.
• Multiply the frequency domain image by the frequency domain filter to obtain a new frequency domain image.
• Perform IDFT on the new frequency domain image to obtain a new image in the spatial domain.

BP神经网络概述

BP神经网络是神经网络中最基础也是应用最多的神经网络，它由三层节点组成，分别为输入层，隐含层和输出层。BP神经网络的实现比较简单，主要分为正向传递输出和反向传递误差两部分。

神经网络的万能逼近定律：一维阶梯函数的线性组合可以逼近任何一维连续函数；sigmoid函数可以逼近阶梯函数，因此一维sigmoid函数的线性组合能够逼近任何连续函数。（1989年，Robert Hecht-Nielsen）这为神经网络的应用提供了理论依据。