| title |
|---|
Using YOLOv5 / YOLOv8 Models for Object Detection in MaixPy |
Object detection refers to identifying the location and category of objects in images or videos, such as detecting apples, airplanes, etc., and marking their positions. Unlike classification, object detection includes positional information, so the result is generally a rectangle marking the object's position.
MaixPy provides YOLOv5 and YOLOv8 models by default, which can be used directly:
YOLOv8 requires MaixPy >= 4.3.0.
from maix import camera, display, image, nn, app
detector = nn.YOLOv5(model="/root/models/yolov5s.mud")
# detector = nn.YOLOv8(model="/root/models/yolov8n.mud")
cam = camera.Camera(detector.input_width(), detector.input_height(), detector.input_format())
dis = display.Display()
while not app.need_exit():
img = cam.read()
objs = detector.detect(img, conf_th = 0.5, iou_th = 0.45)
for obj in objs:
img.draw_rect(obj.x, obj.y, obj.w, obj.h, color = image.COLOR_RED)
msg = f'{detector.labels[obj.class_id]}: {obj.score:.2f}'
img.draw_string(obj.x, obj.y, msg, color = image.COLOR_RED)
dis.show(img)Video demonstration:
Here, the camera captures an image, which is then passed to the detector for detection. The results (classification names and positions) are displayed on the screen.
You can switch between YOLOv5 and YOLOv8 by replacing them in the code. Make sure to update the model file path accordingly.
The list of 80 supported object categories can be found in the appendix.
For more API usage, refer to the documentation of the maix.nn module.
The default model input resolution is 320x224, which closely matches the provided screen resolution. You can manually download models with other resolutions if needed:
YOLOv5: https://maixhub.com/model/zoo/365 YOLOv8: https://maixhub.com/model/zoo/400
Higher resolutions increase accuracy but also require more processing time. Choose the appropriate resolution based on your application.
We provide YOLOv5s and YOLOv8n models. The former is larger and more accurate, while the latter is slightly faster but with marginally lower accuracy. You can test both to see which suits your needs better.
You can also try YOLOv8s, which offers higher accuracy but lower frame rates(e.g. yolov8s_320x224 slower than yolov8n_320x224 about 10ms）. Models can be downloaded from the mentioned model library.
When using the detector.detect(img) function, if img resolution differs from the model's resolution, the function will automatically call img.resize to adjust the image to the model's input resolution. The resize method uses image.Fit.FIT_CONTAIN by default, maintaining the aspect ratio and padding with black. The detected coordinates will map back to the original img coordinates.
If the default 80-class model doesn't meet your needs, visit MaixHub to learn and train your object detection model. Select Object Detection Model when creating a project. Refer to MaixHub online train doc.
Alternatively, check out the models shared by community members in the MaixHub Model Library.
We strongly recommend using MaixHub for online training. Offline training is more complex and not suggested for beginners. This method assumes some pre-existing knowledge not covered in this document. For more details, see Offline Training YOLOv5 Model.
The 80 object categories from the COCO dataset are:
person
bicycle
car
motorcycle
airplane
bus
train
truck
boat
traffic light
fire hydrant
stop sign
parking meter
bench
bird
cat
dog
horse
sheep
cow
elephant
bear
zebra
giraffe
backpack
umbrella
handbag
tie
suitcase
frisbee
skis
snowboard
sports ball
kite
baseball bat
baseball glove
skateboard
surfboard
tennis racket
bottle
wine glass
cup
fork
knife
spoon
bowl
banana
apple
sandwich
orange
broccoli
carrot
hot dog
pizza
donut
cake
chair
couch
potted plant
bed
dining table
toilet
tv
laptop
mouse
remote
keyboard
cell phone
microwave
oven
toaster
sink
refrigerator
book
clock
vase
scissors
teddy bear
hair drier
toothbrush