[PYTHON] ROS Lecture 119 Correct the color of the camera image

environment

This article works in the following environment.

For installation, refer to ROS Course 02 Installation. Also, the program in this article has been uploaded to github. See ROS Course 11 git repository.

Overview

In a previous article, I made an example of detecting an object by hue, but this assumes that the camera has acquired the correct color. However, the colors that can be obtained with the camera will change depending on the performance of the camera and the lighting conditions. Therefore, here is a simple method for color proofing of the camera (and environment). As a general rule, take a picture of an object of "reference color" with a camera, investigate what kind of color it looks like, and find the gain value of each color (red, green, blue). This time, we will use a color checker as the "reference color". A decent one (X-rite ColorChecker Classic) is good, but decent I ordered something similar on Aliexpress because it's expensive. The color checker contains data about colors and how they should appear in RGB.

Source code

color_rectify_tool.py

• It is assumed that the camera image appears as / image_raw.
• When you start it, the screen generated by opencv2 will appear. Click when the color checker is captured in the camera. After that, click the 4 corners of the color checker in the picture to automatically calculate 24 measurement points. If you calculate again, the result will be spit out in HTML.

Run

Run the tool

rosrun cam_lecture color_rectify_tool.py

After execution, the result is as follows. Gian is large = It means that the image is stronger than the original color. Red is weak and blue is strong. This is a typical fluorescent lighting trend.

comment

• This time, all 24 gains were measured and the median average of 12 was adopted. If there is a use, it is better to create correction information centered on the corresponding color.
• Since the correction value has come out, use it by dividing the color of the original image by gain or multiplying the judgment ground of the color judgment by gain.

reference

• [Template Library](https://qiita.com/mima_ita/items/5405109b3b9e2db42332#html%E7%94%A8%E3%81%AE%E3%83%86%E3%83%B3%E3%83% 97% E3% 83% AC% E3% 83% BC% E3% 83% 88% E3% 83% A9% E3% 82% A4% E3% 83% 96% E3% 83% A9% E3% 83% AA)
• [Python hexadecimal display](https://www.it-swarm-ja.tech/ja/python/python%E3%81%A716%E9%80%B2%E6%95%B0%E3%82%92rgb % E5% 80% A4% E3% 81% AB% E5% A4% 89% E6% 8F% 9B% E3% 81% 99% E3% 82% 8B / 105231573 /)
• Display in browser
• color checker

Link to table of contents page

Link to ROS course table of contents