交通信号灯的检测是复杂场景下交通灯识别的重点。采用了色彩分割与关联滤波方案进行交通灯的检测。首先，建立交通信号灯的高斯模型，利用高斯向量与多色彩空间结合的方法进行图像分割。然后，采用基于区域增长与相似性判定的关联滤波，对色彩分割后的图像进行处理。最后，使用基于canny算子的边缘提取算法获取方向指示灯轮廓特征，并使用基于改进hu不变矩和马氏距离对方向指示信号灯进行分类。-The detection of traffic lights is the focus of traffic lights

Using matlab to calculate the Mahalanobis distance for the image, Prediction Error Method for Parameter Identification - the idea of relaxation, You can achieve data classification and regression pattern recognition.

Of spherical harmonics graphic simulation, You can achieve data classification and regression pattern recognition, In the MATLAB image texture feature.

Build a framework OFDM communication system, Including compression ratio, image restoration computing uptime and peak signal to noise ratio, You can achieve data classification and regression pattern recognition.