Grayscale Algorithm: Seven Step Approach

Loading...

来自 Duke University 的课程

Java 程序设计：使用软件解题

2026 个评分

Duke University

Java 程序设计：使用软件解题

2026 个评分

课程 1（共 4 门，Specialization Object Oriented Programming in Java）

Building upon Duke’s “Programming and the Web for Beginners” course, you will grow in your ability to be a problem-solver and programmer by writing Java programs. We are excited that you've chosen to learn Java with us! You will soon be able to solve problems that you find interesting and useful by practicing a design process you will learn in this course. Java concepts presented in this course include basic control flow; object-oriented concepts including classes and methods; custom, open-source classes; and a design-implement-debug process that you can continue to leverage as you build your experience and knowledge of programming principles. The custom, open-source classes will provide a consistent application programming interface (API) so that you can write programs that access and transform images, websites, comma separated value (CSV) files and other data files using simple and accessible coding idioms built on the Java Iterable interface. You will learn to design, test, and debug programs and to find and read Java documentation that can serve as the basis for continued study of Java and programming. After completing this course you will be able to: 1. Edit, compile, and run a Java program; 2. Use conditionals and for loops appropriately in a Java program; 3. Use Java API documentation in writing programs. 4. Debug a Java program using the scientific method; 5. Develop a Java method to solve a specific problem; 6. Develop a set of test cases as part of developing a program; 7. Create a class with multiple methods that work together to solve a problem; and 8. Use divide-and-conquer design techniques for a program that uses more than one method.

从本节课中

MiniProject: Baby Names

This module wraps up the course with a mini project that ties together the different practices, skills, and libraries you have gained across the course! Using data on the popularity of different baby names in the United States from the past several decades, you will be able to compare different names’ popularity over time. While the data we have collected for this course is from the United States, we welcome you to share data from other countries in the course discussion forums. Good luck with the mini project!

Batch Grayscale: Converting Many Files2:58

Grayscale Algorithm: Seven Step Approach4:11

Image Iterable in BlueJ: Grayscale5:36

Batch Processing Grayscale3:30

Saving Images with New Names11:14

Batch Grayscale Summary: Converting Many Files1:09

与讲师见面

Owen Astrachan
Professor of the Practice
Computer Science
Robert Duvall
Lecturer
Computer Science
Andrew D. Hilton
Associate Professor of the Practice
Electrical and Computer Engineering
Susan H. Rodger
Professor of the Practice
Computer Science

Welcome back.

You're currently learning how to convert a bunch of images to gray-scale.

An important component of that process is the ability to convert

one image to gray-scale.

In particular, you want to take a color image and

produce an image that looks like it, but only uses shades of gray, not any colors.

As with all programming problems, the way you want to approach this problem

is with the 7 step approach you learned previously.

The first step is to work an instance by hand.

As always, we must work with a small, manageable problem size.

Here we picked a 2x2 image to work with.

We're going to want to make a 2x2 image for the output, but

how do you figure out what shade of gray to use for this pixel?

You need the domain knowledge before you can proceed.

In this particular problem, the knowledge you need is about colors or graphics.

The first thing we need to know is, what precisely is a shade of gray?

A color is a shade of gray if its red, blue, and

green components are all the same.

However, this knowledge by itself is not sufficient to tell you how to come up

with the shade of gray for a particular color.

Just that the result needs to have their red, blue, and green all the same.

One way you could do this is to average the red, green, and blue components.

Or you might decide you want a weighted average,

because the human eye does not perceive all colors in the same way.

There could be more complex alternatives.

However, just taking the average works pretty well and is simple.

Now you have the domain knowledge required to do this problem yourself.

You can look at the RGB for a pixel, compute the average, and

color in the output pixel appropriately.

Then you would go through looking at the RGB values for each input pixel,

computing their average and coloring in the output pixel accordingly.

Once you have colored in all of the pixels,

you've worked an instance of the problem yourself and are done with Step 1.

The next thing you need to do, is write down exactly what you just did.

I started with the image that I wanted, which we called inImage.

Then, I made another image of the same size, which we called outImage.

I computed (255+0+0)/3, which is 83, and

I made the first pixel of outImage have red, green, and blue values at 83.

And then I went through each other pixel computing the average of the R,

G, and B and coloring the output pixel accordingly.

One we finished this we had these ten steps,

that we used to solve this particular instance of the problem.

Now, you are ready to move to Step 3, and look for patterns and repetition.

You can see that we are doing very similar things to each pixel, but

they are not quite the same.

We need to find the patterns in the numbers to generalize these steps

to any image.

Let us look at the particular numbers we need to generalize.

Why did we use 255 here and 0 here.

These numbers were all the corresponding pixel in inImage's red component.

What about these numbers?

Similarly, these were the green component of the corresponding pixel in inImage.

Lastly, these numbers were the blue component of the corresponding pixel.

Next, you should give a name to the result of this math.

It won't always be these particular numbers, and

you will want to be able to refer to it precisely.

We'll call it average.

Okay, now that we have thought that through,

you can write the general algorithm.

Notice how we thought about what we do for each pixel and wrote down general steps.

Now we can write this in terms of steps to do to each pixel in the output image.

It will work for any size image, with any colors.

The last thing you should do before you write code is to test your generalized

algorithm out on another small input.

Here is a small image and the RGB values for each pixel.

Take a moment to execute the algorithm and see if you get the correct answer.

Yes the answer is right, so you are ready to implement it in code.

探索我们的目录

免费加入并获得个性化推荐、更新和优惠。

Coursera 致力于普及全世界最好的教育，它与全球一流大学和机构合作提供在线课程。

© 2018 Coursera Inc. 保留所有权利。

通过 App Store 下载

通过 Google Play 获取