lstm用于时间序列分析和预测

本程序采用迭代对输入彩色图象进行增强处理,处理速度快,效果好. (this programme is used interactive method to enhance the imput color image)

ANN code to train it

已知一组二维模式矢量如下（20个）,利用k均值聚类算法将模式集分类，并计算出聚类中心。利用输入层为2个神经元，输出层结构为7*7方阵的自组织映射（SOFM）对同样模式进行聚类(Given a set of two-dimensional pattern vector as follows (20), the use of k-means clustering algorithm to classify the pattern set, and calculate the cluster center. Using the input layer 2 neurons, the output layer structure for self-organizing map 7* 7 phalanx (SOFM) for the same model clustering)

Frequency response of second order system

SPWM INVERTR USING RL LOAD

2D version of Brownian motion. A random walk is similar to markov process, however a markov process has no memory, where a random walk uses an initial state. The next step in the walk does not necessarily depend on your current state but will be referenced from your current state, that is to say the random vector is added to your current state. depending on complexity

Flash intro website template

利用MATLAB对原有声音波形进行加躁，去噪。得出加躁语音频谱图、去噪语音频谱图，以及对应文件。 (The original voice waveform with manic using MATLAB, denoising. The impatient voice spectrum, denoising voice spectrum graph, and the corresponding file.)

本仿真做了脉冲压缩雷达的处理过程，包括加窗处理和匹配处理，改仿真代码很精致，希望能够帮助大家；(The simulation pulse compression radar processing, including windowing processing and matching processing)

滑动聚束模式的RD成像。聚束模式SAR中,天线始终指向场景中心,合成孔径的长度比相同条件下的条带式SAR长。对于滑动聚束模式SAR系统,由于天线的指向变化与两者都不同,所以与条带模式相比,其合成孔径更长,而且测绘带区域比聚束模式更大 (The Sliding Spotlight model described in the article can geta better resolution than strip-map mode and better observed area than spotlight mode.by steering the antenna on a point further away the radar than the area beingilluminated which can be regarded as coming the underground. By doing this, thismode can get good combination between the resolution and observed area. )