Python 将依次读入的数据合为一个数据

方案1：
data_path = ['1.jpeg', '2.jpg', '3.jpg', '4.jpg', '5.jpg']

data_all = []
data_all = list(data_all)
for i in range(len(data_path)):
    data = read_image(data_path[i])
    H, W, C = np.shape(data)
    data_all.extend(list(np.resize(data, (H*W, C))))

方案2：

从结尾写入txt，之后再读入txt

with open('data.txt','ab') as f:
    np.savetxt(f, data, delimiter=" ")  # 注意这里的f如果误打成data.txt,则只保留最后保留的，前面保留的都被删除了

读取的时候：
with open('data.txt','r') as f:
    data = np.loadtxt(f)

应用：

将数据集做PCA：

for...

        # 数据读入

        data = np.resize(data_for_pca, (H*W), C)

        data = data[::10, :] # 采样，防止数据规模太大

        with open('data.txt','ab') as f:
             np.savetxt(f, data)  

# 数据集pca计算

with open('data.txt','r') as f:
    data = np.loadtxt(f)

X_norm, means, stds = featureNormalize(pca_data)

U,S,V=pca(X_norm)

Z=self.projectData(pca_data, U, 4)

def featureNormalize(X):

        means = X.means(axis=0)

        stds = X.std(axis=0, ddof=1)

        X_norm=(X-means)/stds

        return X_norm, means, stds

def projectData(X,U,K):

        Z = X@U[:, :K]

        return Z

def pca(X):

        sigma = (X.T@X)/len(X)

        U,S,V = np.linalg.svd(sigma)

        return U, S, V