【深度学习】使用YOLO11和霍夫变换追踪站台黄线穿越者

from ultralytics import YOLOimport argparseimport cv2
if __name__ == '__main__':'''    Apply bbox to detected persons    '''    parser = argparse.ArgumentParser()    parser.add_argument('--image_path', type=str, default="resources/images/frame_yellow_line_0.png")    opt = parser.parse_args()    image_path = opt.image_path
# Load a pretrained YOLO11n model    model = YOLO("yolo11n.pt")    image = cv2.imread(image_path)    results = model.predict(image)
# check that person has index name=0    print("results[0].names: ", results[0


    
].names)
# iter over results. If there is only one frame then results has only one componentfor image_pred in results:        class_names = image_pred.names        boxes = image_pred.boxes
# iter over the detected boxes and select thos of the person if existsfor box in boxes:if class_names[int(box.cls)] == "person":                print("person")                print("person bbox: ", box.xyxy)        image_pred.plot()    image_pred.show()

from ultralytics import YOLOimport cv2font = cv2.FONT_HERSHEY_DUPLEX
# Load a pretrained YOLO11n modelmodel = YOLO("yolo11n.pt")path_image = "resources/images/frame_yellow_line_900.png"image = cv2.imread(path_image)annotated_frame = image.copy()
# set in the predict function the interested classes to detect. Here I want to detect persons, whose index is 0results = model.predict(image, classes=[0], conf=0.54)image_pred = results[0]boxes = image_pred.boxes
# iter over all the detected boxes of personsfor box in boxes:
    x1 = int(box.xyxy[0][0])    y1 = int(box.xyxy[0][1])    x2 = int(box.xyxy[0][2])    y2 = int(box.xyxy[0][3])    coords = (x1, y1 - 10)    text = "person"    print("x1: {} - y1: {} - x2: {} - y2: {}".format(x1, y1, x2, y2))
    color = (0, 255, 0) # colors in BGR    thickness = 3    annotated_frame = cv2.rectangle(image, (x1, y1), (x2, y2), color, thickness)    annotated_frame = cv2.putText(annotated_frame, text, coords, font, 0.7, color, 2)
annotated_frame_path = "/home/enrico/Projects/VideoSurveillance/resources/images/annotated_frame_900.png"cv2.imwrite(annotated_frame_path, annotated_frame)

from ultralytics import YOLOimport cv2font = cv2.FONT_HERSHEY_DUPLEX# Load a pretrained YOLO11n-pose Pose modelmodel = YOLO("yolo11m-pose.pt")# Run inference on an imagepath_image = "resources/images/frame_yellow_line_900.png"image = cv2.imread(path_image)cv2.imwrite(annotated_frame_bbox_path, annotated_frame_bbox)annotated_frame_keypoints = image.copy()annotated_frame_bbox = image.copy()results = model(image)  # results list
# extract keypointskeypoints = results[0].keypointsconf = keypoints.confxy = keypoints.xyprint(xy.shape)  # (N, K, 2) where N is the number of person detectedprint("Detected person: ", xy.shape[0])
# iter over personsfor idx_person in range(xy.shape[0]):
print("idx_person: ", idx_person)
#iter over keypoints of a fixed personlist_x = []list_y = []for i, th in enumerate(xy[idx_person]):x = int(th[0])y = int(th[1])
if x !=0.0 and y!=0.0:
list_x.append(x)list_y.append(y)print("x: {} - y: {}".format(x, y))annotated_frame_keypoints = cv2.circle(annotated_frame_keypoints, (x,y), radius=3, color=(0, 0, 255), thickness=-1)annotated_frame_keypoints = cv2.putText(annotated_frame_keypoints, str(i), (x, y-5), font, 0.7, (0, 0, 255), 2)
if len(list_x) > 0 and len(list_y) > 0:min_x = min(list_x)max_x = max(list_x)min_y = min(list_y)max_y = max(list_y)print("min_x: {} - max_x: {} - min_y: {} - max_y: {}".format(min_x, max_x, min_y, max_y))w = max_x - min_xh = max_y - min_ydx = int(w/3)x0 = min_x - dxx1 = max_x + dxy0 = min_y - dxy1 = max_y + dxprint("x0: {} - x1: {} - y0: {} - y1: {}".format(x0, x1, y0, y1))
coords = (x0, y0 - 10)text = "person"color = (0, 255, 0) # colors in BGRthickness = 3annotated_frame_bbox = cv2.rectangle(annotated_frame_bbox, (x0, y0), (x1, y1), color, thickness)annotated_frame_bbox = cv2.putText(annotated_frame_bbox, text, coords, font, 0.7, color, 2)

annotated_frame_path = "/home/enrico/Projects/VideoSurveillance/resources/images/annotated_frame_keypoints_900.png"cv2.imwrite(annotated_frame_path, annotated_frame_keypoints)
annotated_frame_bbox_path = "/home/enrico/Projects/VideoSurveillance/resources/images/annotated_frame_keypoints_bbox_900.png"

import cv2import numpy as np

deffind_slope(x1, y1, x2, y2):if x2 != x1:return ((y2 - y1) / (x2 - x1))else:return np.inf

deffind_m_and_q(edges):
    lines = cv2.HoughLinesP(            image=edges, # Input edge image            rho=1, # Distance resolution in pixels            theta=np.pi/180, # Angle resolution in radians            threshold=120, # Min number of votes for valid line            minLineLength=80, # Min allowed length of line            maxLineGap=10# Max allowed gap between line for joining them            )
    coefficient_list = []
# Iterate over pointsif lines isNone:# if line is None return an empty list of coefficientsreturn coefficient_listelse:for points in lines:            x_vertical = None
# Extracted points nested in the list            x1,y1,x2,y2=points[0]
            slope = find_slope(x1, y1, x2, y2)if slope == np.inf:# if the slope is infinity the intercept is None and set the x vertical                intercept = None                x_vertical = x1else:                intercept = y1-(x1*(y2-y1)/(x2-x1))
            coefficient_list.append((slope, intercept, x_vertical))    print("coefficient_list: ", coefficient_list)return coefficient_list

defdraw_lines(image, list_coefficient):    image_line = image.copy()


    
    h, w = image_line.shape[:2]for coeff in list_coefficient:        m, q, x_v = coeff        y0 = 0        y1 = hif m != np.inf:            x0 = -q/m            x1 = (h-q)/melse:            x0 = x1 = x_v        cv2.line(image_line, (int(x0), int(y0)), (int(x1), int(y1)), (0, 255, 0), 6)return image_line