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

import argparse
import os
import cv2
import numpy as np
import paddle
from paddle.vision.transforms import Compose, Normalize
from paddledet.utils.logger import setup_logger
from paddledet.models.detectors import YOLOv3
from paddledet.datasets.builder import build_dataset
from paddledet.config import Config
from paddledet.core.workspace import load_config, create
from paddledet.utils.visualizer import draw_boxes
def parse_args():
    parser = argparse.ArgumentParser(description='PaddleDetection model inference')
    parser.add_argument('--model_dir', type=str, default='./output/yolov3_darknet53_270e_coco', help='Model directory path')
    parser.add_argument('--image_path', type=str, default='./test.jpg', help='Path of the input image')
    parser.add_argument('--score_thresh', type=float, default=0.5, help='Threshold of the score')
    parser.add_argument('--output_dir', type=str, default='./output', help='Output directory path')
    parser.add_argument('--use_gpu', type=bool, default=True, help='Use GPU or not')
    args = parser.parse_args()
    return args
def main():
    args = parse_args()
    # 加载配置文件
    cfg_file = os.path.join(args.model_dir, 'pp-yolo.yaml')
    cfg = load_config(cfg_file)
    # 创建数据集
    dataset = build_dataset(cfg.data.test)
    # 创建模型
    model = YOLOv3(
        num_classes=len(dataset.class_names),
        backbone=cfg.model.backbone,
        neck=cfg.model.neck,
        head=cfg.model.head,
        train_cfg=cfg.train_cfg,
        test_cfg=cfg.test_cfg)
    # 加载模型权重
    model_state_dict = paddle.load(os.path.join(args.model_dir, 'model.pdparams'))
    model.set_state_dict(model_state_dict)
    if args.use_gpu:
        paddle.set_device('gpu')
        model.cuda()
    else:
        paddle.set_device('cpu')
    # 预处理图像
    transforms = Compose([
        Normalize(
            mean=cfg.img_mean, std=cfg.img_std, to_rgb=True),
    ])
    img = cv2.imread(args.image_path)
    inputs = transforms(img)
    inputs = np.expand_dims(inputs, axis=0)
    # 进行预测
    model.eval()
    with paddle.no_grad():
        outputs = model.forward(inputs)
    # 处理预测结果
    bboxes = paddle.fluid.layers.multiclass_nms(
        outputs[0],
        outputs[1],
        score_threshold=args.score_thresh,
        nms_top_k=400,
        keep_top_k=100,
        nms_threshold=0.45,
        background_label=-1)
    # 可视化预测结果
    out_img = draw_boxes(img, bboxes, dataset.class_names)
    cv2.imwrite(os.path.join(args.output_dir, 'out.jpg'), out_img)
if __name__ == '__main__':
    main()

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
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226

import codecs
import os
import time
import sys
sys.path.append('PaddleDetection')
import json
import yaml
from functools import reduce
import multiprocessing
from PIL import Image
import cv2
import numpy as np
import paddle
# import paddleseg.transforms as T
from paddle.inference import Config
from paddle.inference import create_predictor
from multiprocessing.dummy import Pool as ThreadPool
from functools import partial
from deploy.python.preprocess import preprocess,Resize, NormalizeImage, Permute, PadStride
from deploy.python.utils import argsparser, Timer, get_current_memory_mb
#id_class_map
LABEL_MAP = {
    "0": "bump",
    "1": "granary",
    "2": "CrossWalk",
    "3": "cone",
    "4": "bridge",
    "5": "pig",
    "6": "tractor",
    "7": "corn",
}
class PredictConfig():
    def __init__(self, model_dir):
        # parsing Yaml config for Preprocess
        deploy_file = os.path.join(model_dir, 'infer_cfg.yml')
        with open(deploy_file) as f:
            yml_conf = yaml.safe_load(f)
        self.arch = yml_conf['arch']
        self.preprocess_infos = yml_conf['Preprocess']
        self.min_subgraph_size = yml_conf['min_subgraph_size']
        self.labels = yml_conf['label_list']
    #     self.print_config()
    # def print_config(self):
    #     print('%s: %s' % ('Model Arch', self.arch))
    #     for op_info in self.preprocess_infos:
    #         print('--%s: %s' % ('transform op', op_info['type']))
def get_test_images(infer_file):
    with open(infer_file, 'r') as f:
        dirs = f.readlines()
    images = []
    for dir in dirs:
        images.append(eval(repr(dir.replace('\n',''))).replace('\\', '/'))
    assert len(images) > 0, "no image found in {}".format(infer_file)
    return images
def load_predictor(model_dir):
    config = Config(
        os.path.join(model_dir, 'model.pdmodel'),
        os.path.join(model_dir, 'model.pdiparams'))
    # initial GPU memory(M), device ID
    config.enable_use_gpu(3000, 0)
    # optimize graph and fuse op
    config.switch_ir_optim(True)
    # disable print log when predict
    config.disable_glog_info()
    # enable shared memory
    config.enable_memory_optim()
    # disable feed, fetch OP, needed by zero_copy_run
    config.switch_use_feed_fetch_ops(False)
    predictor = create_predictor(config)
    return predictor, config
def create_inputs(imgs, im_info):
    inputs = {}
    im_shape = []
    scale_factor = []
    if len(imgs) == 1:
        inputs['image'] = np.array((imgs[0], )).astype('float32')
        inputs['im_shape'] = np.array(
            (im_info[0]['im_shape'], )).astype('float32')
        inputs['scale_factor'] = np.array(
            (im_info[0]['scale_factor'], )).astype('float32')
        return inputs
    for e in im_info:
        im_shape.append(np.array((e['im_shape'], )).astype('float32'))
        scale_factor.append(np.array((e['scale_factor'], )).astype('float32'))
    inputs['im_shape'] = np.concatenate(im_shape, axis=0)
    inputs['scale_factor'] = np.concatenate(scale_factor, axis=0)
    imgs_shape = [[e.shape[1], e.shape[2]] for e in imgs]
    max_shape_h = max([e[0] for e in imgs_shape])
    max_shape_w = max([e[1] for e in imgs_shape])
    padding_imgs = []
    for img in imgs:
        im_c, im_h, im_w = img.shape[:]
        padding_im = np.zeros(
            (im_c, max_shape_h, max_shape_w), dtype=np.float32)
        padding_im[:, :im_h, :im_w] = img
        padding_imgs.append(padding_im)
    inputs['image'] = np.stack(padding_imgs, axis=0)
    return inputs
class Detector(object):
    def __init__(self,
                 pred_config,
                 model_dir):
        self.pred_config = pred_config
        self.predictor, self.config = load_predictor(model_dir)
        self.preprocess_ops = self.get_ops()
        def get_ops(self):
        preprocess_ops = []
        for op_info in self.pred_config.preprocess_infos:
            new_op_info = op_info.copy()
            op_type = new_op_info.pop('type')
            preprocess_ops.append(eval(op_type)(**new_op_info))
        return preprocess_ops
    def predict(self, inputs):
        # preprocess
        input_names = self.predictor.get_input_names()
        for i in range(len(input_names)):
            input_tensor = self.predictor.get_input_handle(input_names[i])
            input_tensor.copy_from_cpu(inputs[input_names[i]])
        # model prediction
        self.predictor.run()
        output_names = self.predictor.get_output_names()
        boxes_tensor = self.predictor.get_output_handle(output_names[0])
        np_boxes = boxes_tensor.copy_to_cpu()
        boxes_num = self.predictor.get_output_handle(output_names[1])
        np_boxes_num = boxes_num.copy_to_cpu()
        # postprocess
        results = []
        if reduce(lambda x, y: x * y, np_boxes.shape) < 6:
            results = {'boxes': np.zeros([]), 'boxes_num': [0]}
        else:
            results = {'boxes': np_boxes, 'boxes_num': np_boxes_num}
        return results
# 将原preprocess的两个参数转为一个参数para
def my_preprocess(para):
    im_path, preprocess_ops = para
    im, im_info = preprocess(im_path, preprocess_ops)
    return im, im_info
def predict_image(detector, image_list, result_path):
    c_results = {"result": []}
    # 不同目标设定不同输出阈值
    multiclass_thres = [0.49, 0.49, 0.49, 0.49, 0.49, 0.49, 0.49, 0.49]
    num_worker = 4
    # processes这个参数可以不设置，如果不设置函数会跟根据计算机的实际情况来决定要运行多少个进程
    pool = ThreadPool(processes=num_worker)# 多线程处理输入图像，预处理速度快一些
    img_length = len(image_list)
    # 根据评估数据自行调整每次多线程处理的样本数量, len(image_list) >= img_iter_filter
    img_iter_filter = 10
    img_iter_range = list(range(img_length//img_iter_filter))
    for start_index in img_iter_range:
        if start_index == img_iter_range[-1]:
            im_paths = image_list[start_index*img_iter_filter:]
        else:
            im_paths = image_list[start_index*img_iter_filter:(start_index+1)*img_iter_filter]
        image_ids = [int(os.path.basename(im_p).split('.')[0]) for im_p in im_paths]
        para = [[i,detector.preprocess_ops] for i in im_paths]
        imandinfos = pool.map(my_preprocess, para)
        # print('imandinfos',imandinfos)
        for idx, imandinfo in enumerate(imandinfos):
            # 检测模型图像预处理
            image_id = image_ids[idx]
            inputs = create_inputs([imandinfo[0]], [imandinfo[1]])
            # 检测模型预测结果
            det_results = detector.predict(inputs)
            # 检测模型写结果
            im_bboxes_num = det_results['boxes_num'][0]
            if im_bboxes_num > 0:
                bbox_results = det_results['boxes'][0:im_bboxes_num, 2:]
                id_results = det_results['boxes'][0:im_bboxes_num, 0]
                score_results = det_results['boxes'][0:im_bboxes_num, 1]
                for idx in range(im_bboxes_num):
                    if float(score_results[idx]) >= multiclass_thres[int(id_results[idx])]:
                        c_results["result"].append({"image_id": image_id,
                                                    "type": LABEL_MAP[str(int(id_results[idx]))],
                                                    "x": float(bbox_results[idx][0]),
                                                    "y": float(bbox_results[idx][1]),
                                                    "width": float(bbox_results[idx][2]) - float(bbox_results[idx][0]),
                                                    "height": float(bbox_results[idx][3]) - float(bbox_results[idx][1]),
                                                    "segmentation": []})
    # 写文件
    with open(result_path, 'w') as ft:
        json.dump(c_results, ft)
def main(infer_txt, result_path, det_model_path):
    pred_config = PredictConfig(det_model_path)
    detector = Detector(pred_config, det_model_path)
    # predict from image
    img_list = get_test_images(infer_txt)
    predict_image(detector, img_list, result_path)
if __name__ == '__main__':
    print('start…')
    start_time = time.time()
    det_model_path = "model/"
    paddle.enable_static()
    infer_txt = sys.argv[1]
    result_path = sys.argv[2]
    main(infer_txt, result_path, det_model_path)
    print('total time:', time.time() - start_time)