Using-color-gamut-limitations-such-as-HSV-and-RGB-for-object-detection

This project summarizes some core traditional visual algorithms commonly used in all drone competitions during undergraduate studies 该项目总结了本科期间参加所有无人机竞赛时常用的一些核心传统视觉算法 A commonly used method for object detection using traditional algorithms is to use color recognition, which can be called color block detection. The specific process is divided into three steps: 传统算法做目标检测有一个很常用的方法是使用颜色识别，可以称为色块检测，具体流程分为三步：

1. Extract the video frame containing the target, and read one or more of the three channel ranges of the target's HSV or RGB color gamut - readhsv.py 1.提取包含目标的视频帧，读取目标的HSV或者RGB等其他色域其中一种或多种的三个通道范围--readhsv.py
2. Further fine tune the range based on the channel range returned in the first step to achieve the best effect -- Fine-tuning_range.py 2.在第一步返回的通道范围基础上进一步微调范围以达到最优效果--Fine_tuning_range.py
3. Use the optimized range after fine-tuning as a prior condition to detect the video stream and return the position of the target - color_track.py 3.使用微调后的最优范围作为先验条件以对视频流做检测，返回目标的位置--color_track.py

Environmental configuration环境配置

pip install opencv
pip install numpy

Code usage代码使用

具体博客在下面链接： 无人机竞赛常用目标检测方法--色块检测

During the use of readhsv.py, the camera window pops up by default. To capture the current frame, you can press the spacebar and then perform operations on the captured video frame. Alternatively, you can use the provided image by pressing the 'q' key. To exit the program finally, you can press the 'esc' key. The key rules for Fine-tuning_range.py are the same. color_track.py uses default thresholds, which can be changed manually. Please note that the program defaults to detecting the built-in image. If you want to use the camera for real-time detection, you can comment out line 20 of the code.

readhsv.py在使用过程中，默认弹出摄像头窗口，如果想要截取当前帧可以按下空格，然后可以在截下的视频帧上进行操作。也可以使用提供的图片，改为按下q键。最终推出程序可以按下esc键。Fine-tuning_range.py按键规则相同。color_track.py使用默认阈值，可以自行更改。注意程序默认对自带图片检测，如果要正常使用摄像头实时检测功能可将第20行代码注释。