object-detection-pcb-errors/pcb_errors.py at master · malgo1311/object-detection-pcb-errors · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
"""
Usage:
  # Make sure you have the folder structure as it is in the github repo

  python pcb_errors.py --input_image=path_to_image

"""

import pandas as pd
import cv2
import numpy as np
import pickle
import sys

import os
# base_folder = "/Users/aishwaryamalgonde/Aishwarya/nanonets"
input_folder = 'inputs'
# input_image_local = '/Users/aishwaryamalgonde/Aishwarya/nanonets/data/2.2.jpg'
# os.chdir(base_folder)

import tensorflow as tf
flags = tf.app.flags
flags.DEFINE_string('input_image', '','Path to the input image file of a pcb')
FLAGS = flags.FLAGS
FLAGS(sys.argv, known_only=True)

positions = pd.read_csv(os.path.join(input_folder,"avg_positions_of_components.csv"))
# positions.head()
with open(os.path.join(input_folder,'missing_threshold_mean.txt'), 'rb') as handle:
    missing_threshold = pickle.loads(handle.read())
# missing_threshold

missing_components_list = ['3_b', '4_b', '33_e', '20_f', '25_f', '24_h', '28_i', '29_i', '30_i',
                           '45_i', '47_n', '48_n', '11_q', '13_s', '42_z', '53_ac', '55_l', '56_ae', '57_af']
rotated_components_list = ['6_6', '20_f', '25_f', '28_i', '30_i', '36_i', '45_i', '44_m',
                            '46_n', '51_n', '47_n', '48_n', '60_o', '13_s', '27_x', '42_z']

common_components_list = []
onlyr_components_list = []
for c in rotated_components_list:
    if c in missing_components_list:
        common_components_list.append(c)
    else:
        onlyr_components_list.append(c)
# common_components_list

onlym_components_list = [x for x in missing_components_list if x not in common_components_list]
# onlym_components_list

test_img = cv2.imread(FLAGS.input_image)
ref_img = cv2.imread(os.path.join(os.getcwd(),input_folder,'ref_img.jpg'))

def load_model(model_path):
    from keras import backend as K
    # This line must be executed before loading Keras model.
    K.set_learning_phase(0)

    from keras.models import load_model
    loaded_model = load_model(model_path)
    return loaded_model

def predict_model(model_path, test_crop):
    model = load_model(model_path)

    img_width = 224
    img_height = 224
    x = cv2.resize(test_crop, (img_width, img_height))
#     x = image.img_to_array(img)
    x = np.expand_dims(x, axis=0)
    x = x/255.

    classes = model.predict(x)
    return classes[0][0]

def main(_):
    error = 0
    ## check for missing components and if they are present then check for rotation
    for c in common_components_list:
        threshold = missing_threshold[c]
        position = positions.loc[positions['component'] == c]

        xmin_avg = int(position['xmin_avg'].mean())
        ymin_avg = int(position['ymin_avg'].mean())
        xmax_avg = int(position['xmax_avg'].mean())
        ymax_avg = int(position['ymax_avg'].mean())

        ref_crop = ref_img[ymin_avg:ymax_avg, xmin_avg:xmax_avg, :]
        test_crop = test_img[ymin_avg:ymax_avg, xmin_avg:xmax_avg, :]

        ref_mean = np.mean(ref_crop, axis = 2)
        test_mean = np.mean(test_crop, axis = 2)
        output = abs(np.mean(ref_mean-test_mean))

        if (output > threshold):
            error += 1
            print('Component', c, 'Missing')
        else:
            model_path = os.path.join(input_folder, 'rotation_detection_models', c, 'vgg16', 'vgg16_best_ld')
            rotation = predict_model(model_path, test_crop)
            if rotation>=0.09:
                error += 1
                print('Component', c.split('_')[0], 'Rotated')

    ## check for missing components only
    for c in onlym_components_list:
        threshold = missing_threshold[c]
        position = positions.loc[positions['component'] == c]

        xmin_avg = int(position['xmin_avg'].mean())
        ymin_avg = int(position['ymin_avg'].mean())
        xmax_avg = int(position['xmax_avg'].mean())
        ymax_avg = int(position['ymax_avg'].mean())

        ref_crop = ref_img[ymin_avg:ymax_avg, xmin_avg:xmax_avg, :]
        test_crop = test_img[ymin_avg:ymax_avg, xmin_avg:xmax_avg, :]

        ref_mean = np.mean(ref_crop, axis = 2)
        test_mean = np.mean(test_crop, axis = 2)
        output = abs(np.mean(ref_mean-test_mean))

        if (output > threshold):
            error += 1
            print('Component', c.split('_')[0], 'Missing')

    ## check for rotation of components olny
    for c in onlyr_components_list:

        position = positions.loc[positions['component'] == c]
        xmin_avg = int(position['xmin_avg'].mean())
        ymin_avg = int(position['ymin_avg'].mean())
        xmax_avg = int(position['xmax_avg'].mean())
        ymax_avg = int(position['ymax_avg'].mean())

        test_crop = test_img[ymin_avg:ymax_avg, xmin_avg:xmax_avg, :]

        model_path = os.path.join(input_folder, 'rotation_detection_models', c, 'vgg16', 'vgg16_best_ld')
        rotation = predict_model(model_path, test_crop)

        if rotation>0.09:
            error += 1
            print('Component', c.split('_')[0], 'Rotated')

    if error == 0:
        print('No Errors')

if __name__ == '__main__':
    tf.app.run()