Lodging grade

 # Automatic grading evaluation of winter wheat lodging based on deep learning

## 1. introduce

In wheat breeding, lodging is a key factor that restricts wheat yield and quality. Timely and accurate classification of winter wheat lodging is of great practical significance for agricultural insurance companies to assess agricultural losses and select good varieties. Measuring lodging classification by lodging degree and lodging area is a commonly used grading method in wheat production. However, in actual production, using artificial field surveys to investigate the tilt angle and lodging area of winter wheat lodging is not only time-consuming and laborious, but the measurement results are subjective and unreliable. In order to address the above situation, this paper proposes a classification semantic segmentation multi task neural network model MLP_U-Net based on an improved MLP module, which can accurately estimate the tilt angle and lodging area of winter wheat lodging, and qualitatively and quantitatively evaluate the grading of wheat lodging.

## 2. config

### config.py

```python

#dataset name You can choose to switch between 1 or 2

#After switching, it needs to be run ceate_dataset.py

datasets_name = '1'

#Switching operating parameters

batch_size = 4

lr = 1e-5

epoch=2000

num_classes=3

num_workers = 1

#Dataset information

train_data = 'cls_train.txt'

test_data = 'cls_test.txt'

classes_path='classes.txt'

#Train log

tr_loss='./train_img/'

val_loss='./val_img/'

test_log='./test_img/'

save_model=r'./model_{}_{}.plt'

save_epoch=200

```

### folder util 

```python

#Data processing related files

```

### train.py

```python

#Used for training datasets

```

### test.py

```python

#Used for testing datasets

```

### test_img/Binarization.py

```python

#Binary the segmentation results

```

### test_img/dele.py

```python

#Clarify the previous test data

```

### test_img/eva.py

```python

#Evaluation Network

```