mirror of
https://github.com/NanjingForestryUniversity/supermachine-tobacco.git
synced 2025-11-08 22:33:54 +00:00
57 lines
2.4 KiB
Python
57 lines
2.4 KiB
Python
import numpy as np
|
||
import scipy
|
||
from imblearn.under_sampling import RandomUnderSampler
|
||
from models import AnonymousColorDetector
|
||
from utils import read_labeled_img
|
||
|
||
# %%
|
||
train_from_existed = False # 是否从现有数据训练，如果是的话，那就从dataset_file训练，否则就用data_dir里头的数据
|
||
data_dir = "data/dataset" # 数据集，文件夹下必须包含`img`和`label`两个文件夹，放置相同文件名的图片和label
|
||
dataset_file = "data/dataset/dataset_2022-07-20_10-04.mat"
|
||
|
||
color_dict = {(0, 0, 255): "yangeng", (255, 0, 0): 'beijing', (0, 255, 0): "zibian"} # 颜色对应的类别
|
||
# color_dict = {(0, 0, 255): "yangeng"}
|
||
# color_dict = {(255, 0, 0): 'beijing'}
|
||
# color_dict = {(0, 255, 0): "zibian"}
|
||
label_index = {"yangeng": 1, "beijing": 0, "zibian": 2} # 类别对应的序号
|
||
show_samples = False # 是否展示样本
|
||
|
||
# 定义一些训练量
|
||
threshold = 5 # 正样本周围多大范围内的还算是正样本
|
||
node_num = 20 # 如果使用ELM作为分类器物，有多少的节点
|
||
negative_sample_num = None # None或者一个数字，对应生成的负样本数量
|
||
# %% md
|
||
## 读取数据
|
||
# %%
|
||
dataset = read_labeled_img(data_dir, color_dict=color_dict, is_ps_color_space=False)
|
||
if show_samples:
|
||
from utils import lab_scatter
|
||
|
||
lab_scatter(dataset, class_max_num=30000, is_3d=True, is_ps_color_space=False)
|
||
# %% md
|
||
## 数据平衡化
|
||
# %%
|
||
if len(dataset) > 1:
|
||
rus = RandomUnderSampler(random_state=0)
|
||
x_list, y_list = np.concatenate([v for k, v in dataset.items()], axis=0).tolist(), \
|
||
np.concatenate([np.ones((v.shape[0],)) * label_index[k] for k, v in dataset.items()],
|
||
axis=0).tolist()
|
||
x_resampled, y_resampled = rus.fit_resample(x_list, y_list)
|
||
dataset = {"inside": np.array(x_resampled)}
|
||
# %% md
|
||
## 模型训练
|
||
# %%
|
||
# 对数据进行预处理
|
||
x = np.concatenate([v for k, v in dataset.items()], axis=0)
|
||
negative_sample_num = int(x.shape[0] * 1.2) if negative_sample_num is None else negative_sample_num
|
||
model = AnonymousColorDetector()
|
||
# %%
|
||
if train_from_existed:
|
||
data = scipy.io.loadmat(dataset_file)
|
||
x, y = data['x'], data['y'].ravel()
|
||
model.fit(x, y=y, is_generate_negative=False, model_selection='dt')
|
||
else:
|
||
world_boundary = np.array([0, 0, 0, 255, 255, 255])
|
||
model.fit(x, world_boundary, threshold, negative_sample_size=negative_sample_num, train_size=0.7,
|
||
is_save_dataset=True, model_selection='dt')
|
||
model.save() |