这种情况下预测是否正确？

Question

请告诉我以下代码的预测是否正确，这是 CNN 的测试代码， 我正在使用 Jupyter、Tensorflow 2.9.0 和 Keras 2.9.0

import math
import os
import numpy as np
import cv2
from keras.models import load_model
from PIL import Image
########### PARAMETERS ##############
threshold = 0.90 # MINIMUM PROBABILITY TO CLASSIFY
#### LOAD THE TRAINNED MODEL 
model = load_model('model_trained.h5')
path = r'C:\Users\Issam\AppData\Local\Programs\Python\Python36\Digits-Classification-master\Weed_Detection\MyTestData\test9.jpg'
# Reading an image in default mode and creating subimages
i = Image.open (path)
width, height = i.size
print(width)
print(height)
L=0
T=0
R=width
B=height
imgOriginal= frame1
img = np.asarray(imgOriginal)
img = cv2.resize(img,(200, 200))
img = preProcessing(img)
cv2.imshow("Processed Image",img)
img = img.reshape(1,200,200,1)
frame1 = i.crop(((L, T, R, B/3)))
def preProcessing(img):
        img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
        img = cv2.equalizeHist(img)
        img = img / 255
        return img
# Predict
predictions = model.predict(img)  
classIndex = np.argmax(predictions,axis=1) 
probVal = np.amax(predictions) 
if probVal > threshold:
    if classIndex == 0:
        item1 = "No_Weed"
        prob1=probVal
        print(item1, "Probability: ",probVal)
    elif classIndex == 1:
        item1 = "Weed"
        prob1=probVal
        print(item1, "Probability: ",probVal)

模型结果如下

Model: "sequential"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 conv2d (Conv2D)             (None, 196, 196, 60)      1560      
                                                                 
 conv2d_1 (Conv2D)           (None, 192, 192, 60)      90060     
                                                                 
 max_pooling2d (MaxPooling2D  (None, 96, 96, 60)       0         
 )                                                               
                                                                 
 conv2d_2 (Conv2D)           (None, 94, 94, 30)        16230     
                                                                 
 conv2d_3 (Conv2D)           (None, 92, 92, 30)        8130      
                                                                 
 max_pooling2d_1 (MaxPooling  (None, 46, 46, 30)       0         
 2D)                                                             
                                                                 
 dropout (Dropout)           (None, 46, 46, 30)        0         
                                                                 
 flatten (Flatten)           (None, 63480)             0         
                                                                 
 dense (Dense)               (None, 500)               31740500  
                                                                 
 dropout_1 (Dropout)         (None, 500)               0         
                                                                 
 dense_1 (Dense)             (None, 2)                 1002      
                                                                 
=================================================================
Total params: 31,857,482
Trainable params: 31,857,482
Non-trainable params: 0
_________________________________________________________________
None

Answer 1

由于我没有足够的积分来发表评论，因此添加评论作为答案。 如果不真正了解模型的架构（尤其是最后一层），就不可能检查模型的输出。您愿意做一个

print(model.summary())

吗？