\n\u26a0\ufe0f \u82e5\u65e0\u8f93\u51fa\uff0c\u8bf7\u5b89\u88c5\u6700\u65b0 NVIDIA \u9a71\u52a8\uff1ahttps:\/\/www.nvidia.com\/Download\/index.aspx<\/p>\n<\/blockquote>\n\n\n\n
\u2705 \u5b89\u88c5 PyTorch \u6240\u9700\u7684 CUDA\uff08\u901a\u8fc7 PyTorch \u5b98\u7f51\uff09<\/h4>\n\n\n\n
\u8bbf\u95ee\uff1ahttps:\/\/pytorch.org\/get-started\/locally\/<\/p>\n\n\n\n
\u9009\u62e9\u547d\u4ee4\uff08\u63a8\u8350 CUDA 11.8 \u6216 12.1\uff09\uff1a<\/p>\n\n\n\n
# \u793a\u4f8b\uff1aCUDA 11.8<\/em>
pip install torch torchvision torchaudio --index-url https:\/\/download.pytorch.org\/whl\/cu118<\/code><\/pre>\n\n\n\n\u9a8c\u8bc1\uff1a<\/p>\n\n\n\n
import torch
print(torch.cuda.is_available())# \u5e94\u4e3a True<\/em>
print(torch.__version__)# \u67e5\u770b\u7248\u672c<\/em><\/code><\/pre>\n\n\n\n
\n\n\n\n2. \u5b89\u88c5 Anaconda<\/h3>\n\n\n\n
\u4e0b\u8f7d\u5730\u5740\uff1ahttps:\/\/www.anaconda.com\/products\/distribution<\/p>\n\n\n\n
\u5b89\u88c5\u540e\u9a8c\u8bc1\uff1a<\/p>\n\n\n\n
conda --version<\/code><\/pre>\n\n\n\n
\n\n\n\n3. \u521b\u5efa Python \u865a\u62df\u73af\u5883<\/h3>\n\n\n\n# \u521b\u5efa\u73af\u5883\uff08Python 3.9 \u63a8\u8350\uff09<\/em>
conda create -n infrared_detection python=3.9
conda activate infrared_detection
# \u5347\u7ea7 pip<\/em>
pip install --upgrade pip<\/code><\/pre>\n\n\n\n
\n\n\n\n4. \u5b89\u88c5\u5fc5\u8981\u4f9d\u8d56<\/h3>\n\n\n\npip install ultralytics opencv-python matplotlib numpy tqdm pillow scikit-learn<\/code><\/pre>\n\n\n\n\n\u2705 ultralytics<\/code>\uff1aYOLOv8 \u5b98\u65b9\u5e93
\u2705 opencv-python<\/code>\uff1a\u56fe\u50cf\u5904\u7406
\u2705 matplotlib<\/code>\uff1a\u53ef\u89c6\u5316
\u2705 tqdm<\/code>\uff1a\u8fdb\u5ea6\u6761<\/p>\n<\/blockquote>\n\n\n\n\u9a8c\u8bc1\u5b89\u88c5\uff1a<\/p>\n\n\n\n
from ultralytics import YOLO
print(\"\u2705 Ultralytics installed!\")<\/code><\/pre>\n\n\n\n
\n\n\n\n\ud83d\udcc1 \u4e8c\u3001\u6570\u636e\u96c6\u7ed3\u6784\u8bbe\u8ba1\uff08\u7ea2\u5916-\u53ef\u89c1\u5149\u5bf9\u9f50\uff09<\/h2>\n\n\n\n\u2705 \u65b9\u6848 A\uff1a\u53cc\u901a\u9053\u8f93\u5165\uff08\u63a8\u8350\u7528\u4e8e\u5bf9\u9f50\u6570\u636e\uff09<\/h3>\n\n\n\n
\u5c06\u7ea2\u5916\u4e0e\u53ef\u89c1\u5149\u62fc\u63a5\u4e3a 3\u901a\u9053\u8f93\u5165\uff08\u5982 R=IR, G=VIS, B=IR\uff09<\/strong> \u6216 4\u901a\u9053\uff08\u9700\u4fee\u6539\u6a21\u578b\uff09<\/strong><\/p>\n\n\n\n\u2705 \u65b9\u6848 B\uff1a\u5206\u522b\u8bad\u7ec3\uff08\u7b80\u5355\u6613\u884c\uff09<\/h3>\n\n\n\n\n- \u5206\u522b\u5728\u7ea2\u5916\u56fe\u50cf\u4e0a\u8bad\u7ec3\u4e00\u4e2a\u6a21\u578b<\/li>\n\n\n\n
- \u5728\u53ef\u89c1\u5149\u56fe\u50cf\u4e0a\u8bad\u7ec3\u53e6\u4e00\u4e2a\u6a21\u578b<\/li>\n\n\n\n
- \u540e\u671f\u878d\u5408\u7ed3\u679c\uff08\u672c\u6559\u7a0b\u91c7\u7528\u6b64\u65b9\u5f0f\u6f14\u793a\uff09<\/li>\n<\/ul>\n\n\n\n
\n\n\n\n\u6570\u636e\u96c6\u7ed3\u6784\uff08YOLO \u683c\u5f0f\uff09<\/h3>\n\n\n\ninfrared_visible_dataset\/
\u251c\u2500\u2500 images\/
\u2502 \u251c\u2500\u2500 train\/
\u2502 \u2502 \u251c\u2500\u2500 ir_0001.jpg
\u2502 \u2502 \u251c\u2500\u2500 vis_0001.jpg
\u2502 \u2502 \u2514\u2500\u2500 ...
\u2502 \u251c\u2500\u2500 val\/
\u2502 \u2514\u2500\u2500 test\/
\u251c\u2500\u2500 labels\/
\u2502 \u251c\u2500\u2500 train\/
\u2502 \u2502 \u251c\u2500\u2500 ir_0001.txt
\u2502 \u2502 \u251c\u2500\u2500 vis_0001.txt
\u2502 \u2502 \u2514\u2500\u2500 ...
\u2502 \u251c\u2500\u2500 val\/
\u2502 \u2514\u2500\u2500 test\/
\u251c\u2500\u2500 data_ir.yaml
\u251c\u2500\u2500 data_vis.yaml<\/code><\/pre>\n\n\n\ndata_ir.yaml<\/code>\uff08\u7ea2\u5916\u6570\u636e\uff09<\/h3>\n\n\n\ntrain: .\/infrared_visible_dataset\/images\/train_ir
val: .\/infrared_visible_dataset\/images\/val_ir
test: .\/infrared_visible_dataset\/images\/test_ir
nc:4
names:['Plane','Drone','Heli','Bird']<\/code><\/pre>\n\n\n\ndata_vis.yaml<\/code>\uff08\u53ef\u89c1\u5149\u6570\u636e\uff09<\/h3>\n\n\n\ntrain: .\/infrared_visible_dataset\/images\/train_vis
val: .\/infrared_visible_dataset\/images\/val_vis
test: .\/infrared_visible_dataset\/images\/test_vis
nc:4
names:['Plane','Drone','Heli','Bird']<\/code><\/pre>\n\n\n\n\n\ud83d\udca1 \u8bf4\u660e\uff1a<\/p>\n\n\n\n
\ntrain_ir<\/code>,\u00a0val_ir<\/code>\uff1a\u53ea\u5305\u542b\u7ea2\u5916\u56fe\u50cf<\/li>\n\n\n\ntrain_vis<\/code>,\u00a0val_vis<\/code>\uff1a\u53ea\u5305\u542b\u53ef\u89c1\u5149\u56fe\u50cf<\/li>\n\n\n\n.txt<\/code>\u00a0\u6587\u4ef6\u683c\u5f0f\u4e3a YOLO\uff1aclass_id x_center y_center width height<\/code>\uff08\u5f52\u4e00\u5316\uff09<\/li>\n<\/ul>\n<\/blockquote>\n\n\n\n
\n\n\n\n\ud83d\ude80 \u4e09\u3001\u8c03\u7528\u5b98\u65b9\u9884\u8bad\u7ec3\u6a21\u578b\u5e76\u8bad\u7ec3\uff08YOLOv8\uff09<\/h2>\n\n\n\n1. \u8bad\u7ec3\u7ea2\u5916\u56fe\u50cf\u6a21\u578b\uff08train_ir.py\uff09<\/h3>\n\n\n\n# train_ir.py<\/em>
from ultralytics import YOLO
import torch
deftrain_infrared_model():
# \u68c0\u67e5\u8bbe\u5907<\/em>
device =0if torch.cuda.is_available()else'cpu'
print(f\"Using device: {device}\")
# \u52a0\u8f7d\u5b98\u65b9\u9884\u8bad\u7ec3\u6a21\u578b\uff08\u63a8\u8350 yolov8s \u6216 yolov8m\uff09<\/em>
model = YOLO('yolov8s.pt')# \u81ea\u52a8\u4e0b\u8f7d<\/em>
# \u5f00\u59cb\u8bad\u7ec3<\/em>
results = model.train(
data='infrared_visible_dataset\/data_ir.yaml',
epochs=150,
imgsz=640,
batch=16,
name='ir_yolov8s',
project='runs\/detect',
save=True,
save_period=10,
device=device,
workers=4,
patience=30,
optimizer='AdamW',
lr0=0.001,
augment=True,
hsv_h=0.015,
hsv_s=0.7,
hsv_v=0.4,
degrees=10.0,
translate=0.2,
scale=0.5,
flipud=0.0,
fliplr=0.5,
mosaic=1.0,
mixup=0.1
)
print(\"\u2705 \u7ea2\u5916\u6a21\u578b\u8bad\u7ec3\u5b8c\u6210\uff01\")
return results
if __name__ =='__main__':
train_infrared_model()<\/code><\/pre>\n\n\n\n2. \u8bad\u7ec3\u53ef\u89c1\u5149\u6a21\u578b\uff08train_vis.py\uff09<\/h3>\n\n\n\n# train_vis.py<\/em>
from ultralytics import YOLO
import torch
deftrain_visible_model():
device =0if torch.cuda.is_available()else'cpu'
model = YOLO('yolov8s.pt')
results = model.train(
data='infrared_visible_dataset\/data_vis.yaml',
epochs=150,
imgsz=640,
batch=16,
name='vis_yolov8s',
project='runs\/detect',
save=True,
device=device,
workers=4,
patience=30,
lr0=0.001
)
print(\"\u2705 \u53ef\u89c1\u5149\u6a21\u578b\u8bad\u7ec3\u5b8c\u6210\uff01\")
return results
if __name__ =='__main__':
train_visible_model()<\/code><\/pre>\n\n\n\n\n\ud83d\udce6 \u6a21\u578b\u4fdd\u5b58\u8def\u5f84\uff1a<\/p>\n\n\n\n
\n- \u7ea2\u5916\uff1a
runs\/detect\/ir_yolov8s\/weights\/best.pt<\/code><\/li>\n\n\n\n- \u53ef\u89c1\u5149\uff1a
runs\/detect\/vis_yolov8s\/weights\/best.pt<\/code><\/li>\n<\/ul>\n<\/blockquote>\n\n\n\n
\n\n\n\n\ud83d\udd0d \u56db\u3001\u63a8\u7406\u4ee3\u7801\uff08detect.py\uff09<\/h2>\n\n\n\n\u5355\u56fe\u63a8\u7406\uff08\u652f\u6301\u53cc\u6a21\u6001\uff09<\/h3>\n\n\n\n# detect.py<\/em>
from ultralytics import YOLO
import cv2
import numpy as np
classMultiModalDetector:
def__init__(self, ir_model_path='runs\/detect\/ir_yolov8s\/weights\/best.pt',
vis_model_path='runs\/detect\/vis_yolov8s\/weights\/best.pt'):
self.ir_model = YOLO(ir_model_path)
self.vis_model = YOLO(vis_model_path)
defdetect(self, ir_image_path, vis_image_path, conf=0.4, output_path='fusion_result.jpg'):
# \u7ea2\u5916\u68c0\u6d4b<\/em>
ir_results = self.ir_model.predict(source=ir_image_path, conf=conf, verbose=False)
ir_img = ir_results[0].plot()
# \u53ef\u89c1\u5149\u68c0\u6d4b<\/em>
vis_results = self.vis_model.predict(source=vis_image_path, conf=conf, verbose=False)
vis_img = vis_results[0].plot()
# \u62fc\u63a5\u5bf9\u6bd4\u56fe<\/em>
combined = np.hstack([ir_img, vis_img])
cv2.imwrite(output_path, combined)
print(f\"\u2705 \u68c0\u6d4b\u5b8c\u6210\uff0c\u7ed3\u679c\u4fdd\u5b58\u81f3: {output_path}\")
# \u8fd4\u56de\u539f\u59cb\u68c0\u6d4b\u7ed3\u679c\uff08\u53ef\u7528\u4e8e\u878d\u5408\uff09<\/em>
return ir_results[0], vis_results[0]
# \u4f7f\u7528\u793a\u4f8b<\/em>
if __name__ =='__main__':
detector = MultiModalDetector()
ir_result, vis_result = detector.detect(
ir_image_path='test_ir.jpg',
vis_image_path='test_vis.jpg'
)<\/code><\/pre>\n\n\n\n
\n\n\n\n\ud83d\udcca \u4e94\u3001\u8bc4\u4f30\u4ee3\u7801\uff08evaluate.py\uff09<\/h2>\n\n\n\n# evaluate.py<\/em>
from ultralytics import YOLO
defevaluate_model(model_path, data_yaml):
model = YOLO(model_path)
metrics = model.val(
data=data_yaml,
batch=16,
imgsz=640,
device=0if torch.cuda.is_available()else'cpu',
save_json=True,
conf_thres=0.001,
iou_thres=0.6
)
print(f\"\u2705 mAP@0.5: {metrics.box.map50:.4f}\")
print(f\"\u2705 mAP@0.5:0.95: {metrics.box.map:.4f}\")
print(f\"Precision: {metrics.box.p:.4f}\")
print(f\"Recall: {metrics.box.r:.4f}\")
print(\"\\n\ud83d\udcca \u5404\u7c7b\u522b AP@0.5:\")
for i, name inenumerate(metrics.names.values()):
print(f\" {name}: {metrics.box.ap[i]:.4f}\")
# \u8bc4\u4f30\u7ea2\u5916\u6a21\u578b<\/em>
print(\"\ud83d\udd0d \u7ea2\u5916\u6a21\u578b\u8bc4\u4f30:\")
evaluate_model('runs\/detect\/ir_yolov8s\/weights\/best.pt','infrared_visible_dataset\/data_ir.yaml')
# \u8bc4\u4f30\u53ef\u89c1\u5149\u6a21\u578b<\/em>
print(\"\\n\ud83d\udd0d \u53ef\u89c1\u5149\u6a21\u578b\u8bc4\u4f30:\")
evaluate_model('runs\/detect\/vis_yolov8s\/weights\/best.pt','infrared_visible_dataset\/data_vis.yaml')<\/code><\/pre>\n\n\n\n
\n\n\n\n\ud83e\udde0 \u516d\u3001\u9ad8\u7ea7\u529f\u80fd\u5efa\u8bae\uff08\u53ef\u9009\uff09<\/h2>\n\n\n\n\u2705 1. \u591a\u6a21\u6001\u878d\u5408\u7b56\u7565\uff08\u540e\u671f\u878d\u5408\uff09<\/h3>\n\n\n\ndeffuse_results(ir_result, vis_result, iou_thres=0.5):
from ultralytics.engine.results import Results
# \u4f7f\u7528 NMS \u6216 Soft-NMS \u878d\u5408\u4e24\u4e2a\u68c0\u6d4b\u7ed3\u679c<\/em>
# \u53ef\u57fa\u4e8e\u7f6e\u4fe1\u5ea6\u52a0\u6743\u5e73\u5747<\/em>
pass<\/code><\/pre>\n\n\n\n\u2705 2. \u5bfc\u51fa\u4e3a ONNX\/TensorRT\uff08\u52a0\u901f\u90e8\u7f72\uff09<\/h3>\n\n\n\nmodel = YOLO('runs\/detect\/ir_yolov8s\/weights\/best.pt')
model.export(format='onnx', dynamic=True, opset=13)
model.export(format='engine', device=0)# TensorRT<\/em><\/code><\/pre>\n\n\n\n\u2705 3. \u6784\u5efa Web \u7cfb\u7edf\uff08Flask\uff09<\/h3>\n\n\n\nfrom flask import Flask, request, send_file
app = Flask(__name__)
detector = MultiModalDetector()
@app.route('\/detect', methods=['POST'])
defrun_detection():
ir = request.files['ir']
vis = request.files['vis']
ir_path ='temp_ir.jpg'
vis_path ='temp_vis.jpg'
ir.save(ir_path)
vis.save(vis_path)
detector.detect(ir_path, vis_path, output_path='result.jpg')
return send_file('result.jpg', mimetype='image\/jpeg')<\/code><\/pre>\n\n\n\n
\n\n\n\n\u4ee5\u4e0a\u6587\u5b57\u53ca\u4ee3\u7801\u4ec5\u4f9b\u53c2\u8003\u5b66\u4e60\u4f7f\u7528\u3002<\/p>\n","raw":"","protected":false},"excerpt":{"rendered":"
\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\u76ee\u6807\u68c0\u6d4b\u7b97\u6cd5YOLOV8\u6a21\u578b\u5982\u4f55\u8bad\u7ec3 \u7ea2\u5916-\u53ef\u89c1\u5149\u5bf9\u9f50\u76ee\u6807\u68c0\u6d4b\u6570\u636e\u96c6 \u7ea2\u5916\u76ee\u6807\u68c0\u6d4b\u6570\u636e\u96c6\u7684\u8bad\u7ec3\u53ca\u5e94\u7528 \u6587\u7ae0\u76ee\u5f55 \u6570\u636e\u96c6 …<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"emotion":"","emotion_color":"","title_style":"","license":"","footnotes":""},"categories":[46],"tags":[],"class_list":["post-6590","post","type-post","status-publish","format-standard","hentry","category-yolo"],"post_thumbnail_image":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","content_first_image":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","post_medium_image_300":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","post_thumbnail_image_624":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","post_frist_image":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","post_medium_image":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","post_large_image":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","post_full_image":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","post_all_images":[{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","id":"image0"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKDgQnibUoltZOwbedyTCHwDtf6UrDow6SibUdcA1j2kGjgTw2g8G6cMicQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=0","id":"image1"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKibEicc0eLjWbBHoibDVBhRTonf4E4W6RzlZ5Nw358icQsUDQDY03PgbTPQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=1","id":"image2"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKibEicc0eLjWbBHoibDVBhRTonf4E4W6RzlZ5Nw358icQsUDQDY03PgbTPQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=1","id":"image3"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKibv1lCr04D2lAajDKXdJrSo4ibiaIAlL3xicMoQwn64HvnlHbMk1ngEE7g\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=2","id":"image4"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKibv1lCr04D2lAajDKXdJrSo4ibiaIAlL3xicMoQwn64HvnlHbMk1ngEE7g\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=2","id":"image5"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKeJFicpzNOem480fGjLOXafBomuS4Vco8BRiciafYUt9plsMiaFPMLhhGYQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=3","id":"image6"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKeJFicpzNOem480fGjLOXafBomuS4Vco8BRiciafYUt9plsMiaFPMLhhGYQ\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=3","id":"image7"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKrOctPEYXLiaLX3QzibWxyF6qExjAZPSdwUE2WMuwca1IQibvOa3xt52Ew\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=4","id":"image8"},{"imagesurl":"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/3aq7xqHVYu3lOF8mVaLAO3D2XLuLXDCKrOctPEYXLiaLX3QzibWxyF6qExjAZPSdwUE2WMuwca1IQibvOa3xt52Ew\/640?wx_fmt=png&from=appmsg&watermark=1&tp=wxpic&wxfrom=5&wx_lazy=1#imgIndex=4","id":"image9"}],"videoAdId":"","listAd":"0","listAdId":"","listAdEvery":6,"total_comments":0,"category_name":"YOLO","post_date":"2025-09-29","like_count":"0","praiseWord":"\u9f13\u52b1","copyright_state":"","excitationAd":"0","rewardedVideoAdId":"","detailAdId":"","detailAd":"0","enterpriseMinapp":"0","audios":[],"postImageUrl":"https:\/\/wp-moto-1258805347.cos.ap-shanghai.myqcloud.com\/2023\/05\/20230519082947553.jpg","avatarurls":[],"related_posts":null,"pageviews":402,"next_post_id":6592,"next_post_title":"YOLOv8\u6846\u67b6\u8bad\u7ec3\u7ea2\u5916\u56fe\u50cf\u5c0f\u76ee\u6807\u68c0\u6d4b\u6570\u636e\u96c6 \u7ea2\u5916\u56fe\u50cf\u5c0f\u76ee\u6807\u68c0\u6d4b\u6570\u636e\u96c6 \u5efa\u7acb\u7ea2\u5916\u5c0f\u76ee\u6807\u68c0\u6d4b\u7cfb\u7edf\u8bc6\u522b\u68c0\u6d4b\u98de\u673a \u65e0\u4eba\u673a \u76f4\u5347\u673a \u9e1f\u7c7b","previous_post_id":6584,"previous_post_title":"YOLOV8\u8bc6\u522b\u5c0f\u6837\u672c\u76ee\u6807\u68c0\u6d4b\u3001\u7a7a\u4e2d\u76ee\u6807\u8bc6\u522b\u975e\u6cd5\u5165\u4fb5\u7684\u65e0\u4eba\u673a \u7a7a\u4e2d\u98de\u884c\u7269\uff0c\u9e1f \u76f4\u5347\u673a \u98de\u884c\u7269\u68c0\u6d4b-\u65e0\u4eba\u673a\u5c0f\u76ee\u6807\u68c0\u6d4b\u6570\u636e\u96c6","_links":{"self":[{"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/posts\/6590","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/comments?post=6590"}],"version-history":[{"count":1,"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/posts\/6590\/revisions"}],"predecessor-version":[{"id":6591,"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/posts\/6590\/revisions\/6591"}],"wp:attachment":[{"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/media?parent=6590"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/categories?post=6590"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/xinchewhd.com.cn\/index.php\/wp-json\/wp\/v2\/tags?post=6590"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
![](\"https:\/\/s.nmxc.ltd\/sakurairo_vision\/@2.6\/load_svg\/inload.svg\")