solarplant / app.py
npc0's picture
Update app.py
af5339f
import os
# import yolov5
# # load model
# model = yolov5.load('keremberke/yolov5m-license-plate')
# # set model parameters
# model.conf = 0.5 # NMS confidence threshold
# model.iou = 0.25 # NMS IoU threshold
# model.agnostic = False # NMS class-agnostic
# model.multi_label = False # NMS multiple labels per box
# model.max_det = 1000 # maximum number of detections per image
# # set image
# def license_plate_detect(img):
# # perform inference
# results = model(img, size=640)
# # inference with test time augmentation
# results = model(img, augment=True)
# # parse results
# if len(results.pred):
# predictions = results.pred[0]
# boxes = predictions[:, :4] # x1, y1, x2, y2
# scores = predictions[:, 4]
# categories = predictions[:, 5]
# return boxes
# from PIL import Image
# # image = Image.open(img)
# import pytesseract
# def read_license_number(img):
# boxes = license_plate_detect(img)
# if boxes:
# return [pytesseract.image_to_string(
# image.crop(bbox.tolist()))
# for bbox in boxes]
from transformers import CLIPProcessor, CLIPModel
vit_model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
def zero_shot_classification(image, labels):
inputs = processor(text=labels,
images=image,
return_tensors="pt",
padding=True)
outputs = vit_model(**inputs)
logits_per_image = outputs.logits_per_image # this is the image-text similarity score
return logits_per_image.softmax(dim=1) # we can take the softmax to get the label probabilities
# installed_list = []
# # image = Image.open(requests.get(url, stream=True).raw)
# def check_solarplant_installed_by_license(license_number_list):
# if len(installed_list):
# return [license_number in installed_list
# for license_number in license_number_list]
def check_solarplant_installed_by_image(image, output_label=False):
zero_shot_class_labels = ["bus with solar panel grids",
"bus without solar panel grids"]
probs = zero_shot_classification(image, zero_shot_class_labels)
if output_label:
return zero_shot_class_labels[probs.argmax().item()]
return probs.argmax().item() == 0
# def check_solarplant_broken(image):
# zero_shot_class_labels = ["white broken solar panel",
# "normal black solar panel grids"]
# probs = zero_shot_classification(image, zero_shot_class_labels)
# idx = probs.argmax().item()
# return zero_shot_class_labels[idx].split(" ")[1-idx]
from fastsam import FastSAM, FastSAMPrompt
os.system('wget https://huggingface.co/spaces/An-619/FastSAM/resolve/main/weights/FastSAM.pt')
model = FastSAM('./FastSAM.pt')
DEVICE = 'cpu'
def segment_solar_panel(img):
# os.system('python Inference.py --model_path FastSAM.pt --img_path bus.jpg --text_prompt "solar panel grids"')
img = img.convert("RGB")
everything_results = model(img, device=DEVICE, retina_masks=True, imgsz=1024, conf=0.4, iou=0.9,)
prompt_process = FastSAMPrompt(img, everything_results, device=DEVICE)
# everything prompt
ann = prompt_process.everything_prompt()
# bbox default shape [0,0,0,0] -> [x1,y1,x2,y2]
ann = prompt_process.box_prompt(bbox=[[200, 200, 300, 300]])
# text prompt
ann = prompt_process.text_prompt(text='solar panel grids')
# point prompt
# points default [[0,0]] [[x1,y1],[x2,y2]]
# point_label default [0] [1,0] 0:background, 1:foreground
ann = prompt_process.point_prompt(points=[[620, 360]], pointlabel=[1])
prompt_process.plot(annotations=ann,output_path='./bus.jpg',)
return Image.Open('./bus.jpg')
import gradio as gr
def greet(img):
if check_solarplant_installed_by_image(img):
seg = segment_solar_panel(img)
return (seg, 'ε˜—θ©¦εˆ†ε‰²ε€ͺι™½θƒ½ζΏιƒ¨εˆ†')
# return (seg,
# "θ»Šη‰ŒοΌš " + '; '.join(lns) + "\n\n" \
# + "ι‘žεž‹οΌš "+ check_solarplant_installed_by_image(img, True) + "\n\n" \
# + "η‹€ζ…‹οΌš" + check_solarplant_broken(img))
return (img, "ζ²’ζœ‰ε€ͺι™½θƒ½ζΏιƒ¨εˆ†εˆ†ε‰²")
iface = gr.Interface(fn=greet, inputs="image", outputs=["image", "text"])
iface.launch()