FootBall-route-drawing-and-modeling/backend/app.py

import os
import uuid
import time
from flask import Flask, request, jsonify, send_from_directory, Response
from flask_cors import CORS
import cv2
import numpy as np
import json
import tempfile
from werkzeug.utils import secure_filename
import threading

from models.player_detector import PlayerDetector
from models.player_tracker import PlayerTracker
from utils.field_transformer import FieldTransformer

app = Flask(__name__, static_folder='uploads')
CORS(app)

# 配置上传文件夹
UPLOAD_FOLDER = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'uploads')
os.makedirs(UPLOAD_FOLDER, exist_ok=True)
os.makedirs(os.path.join(UPLOAD_FOLDER, 'videos'), exist_ok=True)
os.makedirs(os.path.join(UPLOAD_FOLDER, 'results'), exist_ok=True)
os.makedirs(os.path.join(UPLOAD_FOLDER, 'frames'), exist_ok=True)  # 存储视频帧
os.makedirs(os.path.join(UPLOAD_FOLDER, 'visualizations'), exist_ok=True)  # 存储可视化结果

# 初始化模型
player_detector = PlayerDetector(confidence=0.6)  # 提高置信度阈值
player_tracker = PlayerTracker()
field_transformer = FieldTransformer()

# 存储处理任务状态
processing_tasks = {}

@app.route('/api/upload', methods=['POST'])
def upload_file():
    if 'video' not in request.files:
        return jsonify({'error': '没有上传文件'}), 400
    
    file = request.files['video']
    if file.filename == '':
        return jsonify({'error': '没有选择文件'}), 400
    
    # 生成唯一ID并保存视频
    video_id = str(uuid.uuid4())
    filename = secure_filename(file.filename)
    video_path = os.path.join(UPLOAD_FOLDER, 'videos', f"{video_id}_{filename}")
    file.save(video_path)
    
    # 创建处理任务
    processing_tasks[video_id] = {
        'status': 'processing',
        'progress': 0,
        'video_path': video_path,
        'filename': filename
    }
    
    # 异步处理视频分析
    threading.Thread(target=process_video_async, args=(video_path, video_id)).start()
    
    return jsonify({
        'success': True,
        'videoId': video_id,
        'message': '视频上传成功，正在处理中',
        'status': 'processing'
    })

def process_video_async(video_path, video_id):
    """异步处理视频"""
    try:
        result = process_video(video_path, video_id)
        processing_tasks[video_id]['status'] = 'completed'
        processing_tasks[video_id]['result'] = result
    except Exception as e:
        processing_tasks[video_id]['status'] = 'error'
        processing_tasks[video_id]['error'] = str(e)
        print(f"处理视频时出错: {str(e)}")

def process_video(video_path, video_id):
    """处理视频并提取球员轨迹"""
    # 读取视频
    cap = cv2.VideoCapture(video_path)
    
    # 获取视频基本信息
    fps = cap.get(cv2.CAP_PROP_FPS)
    total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
    
    # 创建帧文件夹
    frames_dir = os.path.join(UPLOAD_FOLDER, 'frames', video_id)
    os.makedirs(frames_dir, exist_ok=True)
    
    # 创建可视化文件夹
    vis_dir = os.path.join(UPLOAD_FOLDER, 'visualizations', video_id)
    os.makedirs(vis_dir, exist_ok=True)
    
    # 存储轨迹数据
    all_tracks = []
    frame_count = 0
    
    # 第一帧用于检测场地（可选）
    ret, first_frame = cap.read()
    if ret:
        # 尝试自动检测场地
        field_detected = field_transformer.auto_detect_field(first_frame)
        
        # 如果自动检测失败，使用默认值
        if not field_detected:
            print("场地自动检测失败，使用默认变换")
    
    # 重置视频读取
    cap.set(cv2.CAP_PROP_POS_FRAMES, 0)
    
    while cap.isOpened():
        ret, frame = cap.read()
        if not ret:
            break
        
        # 更新进度
        if video_id in processing_tasks:
            processing_tasks[video_id]['progress'] = min(95, int((frame_count / total_frames) * 100))
        
        # 保存帧
        frame_path = os.path.join(frames_dir, f"frame_{frame_count:06d}.jpg")
        cv2.imwrite(frame_path, frame)
        
        # 检测球员
        detections = player_detector.detect(frame)
        
        # 跟踪球员
        tracks = player_tracker.update(detections, frame)
        
        # 变换坐标到标准足球场
        field_positions = field_transformer.transform_to_field(tracks, frame)
        
        # 可视化跟踪结果
        vis_frame = player_tracker.visualize(frame, draw_history=True)
        vis_path = os.path.join(vis_dir, f"vis_{frame_count:06d}.jpg")
        cv2.imwrite(vis_path, vis_frame)
        
        # 保存当前帧的轨迹数据
        frame_data = {
            'frame': frame_count,
            'tracks': []
        }
        
        for track_id, position in field_positions.items():
            # 从原始跟踪数据获取球员位置和尺寸
            if track_id in tracks:
                x, y, w, h = tracks[track_id]
                x1, y1 = int(x - w/2), int(y - h/2)
                x2, y2 = int(x + w/2), int(y + h/2)
                
                frame_data['tracks'].append({
                    'id': int(track_id),
                    'x': float(position[0]),
                    'y': float(position[1]),
                    'z': 0.0,  # 高度默认为0
                    'image_pos': {
                        'x1': x1,
                        'y1': y1,
                        'x2': x2,
                        'y2': y2
                    }
                })
        
        all_tracks.append(frame_data)
        frame_count += 1
    
    cap.release()
    
    # 保存结果到JSON文件
    result_path = os.path.join(UPLOAD_FOLDER, 'results', f"{video_id}_tracks.json")
    with open(result_path, 'w') as f:
        json.dump(all_tracks, f)
    
    # 保存视频信息
    video_info = {
        'fps': fps,
        'total_frames': total_frames,
        'width': width,
        'height': height,
        'filename': os.path.basename(video_path)
    }
    
    video_info_path = os.path.join(UPLOAD_FOLDER, 'results', f"{video_id}_info.json")
    with open(video_info_path, 'w') as f:
        json.dump(video_info, f)
    
    return {
        'trackFile': f"/api/results/{video_id}_tracks.json",
        'infoFile': f"/api/results/{video_id}_info.json",
        'totalFrames': frame_count,
        'framesDir': f"/api/frames/{video_id}",
        'visualizationsDir': f"/api/visualizations/{video_id}",
        'fps': fps
    }

@app.route('/api/status/<video_id>', methods=['GET'])
def get_status(video_id):
    """获取视频处理状态"""
    if video_id not in processing_tasks:
        return jsonify({'error': '找不到处理任务'}), 404
    
    task = processing_tasks[video_id]
    response = {
        'status': task['status'],
        'progress': task['progress']
    }
    
    if task['status'] == 'completed' and 'result' in task:
        response['result'] = task['result']
    elif task['status'] == 'error' and 'error' in task:
        response['error'] = task['error']
    
    return jsonify(response)

@app.route('/api/results/<filename>', methods=['GET'])
def get_result(filename):
    return send_from_directory(os.path.join(UPLOAD_FOLDER, 'results'), filename)

@app.route('/api/frames/<video_id>/<frame_filename>', methods=['GET'])
def get_frame(video_id, frame_filename):
    return send_from_directory(os.path.join(UPLOAD_FOLDER, 'frames', video_id), frame_filename)

@app.route('/api/visualizations/<video_id>/<vis_filename>', methods=['GET'])
def get_visualization(video_id, vis_filename):
    return send_from_directory(os.path.join(UPLOAD_FOLDER, 'visualizations', video_id), vis_filename)

@app.route('/api/video/<video_id>', methods=['GET'])
def get_video(video_id):
    """获取原始视频"""
    # 查找对应的视频文件
    for filename in os.listdir(os.path.join(UPLOAD_FOLDER, 'videos')):
        if filename.startswith(video_id):
            return send_from_directory(os.path.join(UPLOAD_FOLDER, 'videos'), filename)
    
    return jsonify({'error': '找不到视频文件'}), 404

@app.route('/api/generate_visualization_video/<video_id>', methods=['GET'])
def generate_visualization_video(video_id):
    """生成包含轨迹可视化的视频"""
    if video_id not in processing_tasks or processing_tasks[video_id]['status'] != 'completed':
        return jsonify({'error': '视频尚未处理完成'}), 400
    
    vis_dir = os.path.join(UPLOAD_FOLDER, 'visualizations', video_id)
    if not os.path.exists(vis_dir) or len(os.listdir(vis_dir)) == 0:
        return jsonify({'error': '找不到可视化结果'}), 404
    
    try:
        # 获取视频信息
        with open(os.path.join(UPLOAD_FOLDER, 'results', f"{video_id}_info.json"), 'r') as f:
            video_info = json.load(f)
        
        fps = video_info.get('fps', 30)
        width = video_info.get('width', 1280)
        height = video_info.get('height', 720)
        
        # 输出视频路径
        output_path = os.path.join(UPLOAD_FOLDER, 'videos', f"{video_id}_visualization.mp4")
        
        # 获取所有可视化帧文件
        frame_files = sorted([f for f in os.listdir(vis_dir) if f.startswith('vis_')])
        
        if not frame_files:
            return jsonify({'error': '没有找到可视化帧'}), 404
        
        # 读取第一帧获取尺寸
        first_frame = cv2.imread(os.path.join(vis_dir, frame_files[0]))
        h, w, _ = first_frame.shape
        
        # 创建视频写入器
        fourcc = cv2.VideoWriter_fourcc(*'mp4v')
        out = cv2.VideoWriter(output_path, fourcc, fps, (w, h))
        
        # 写入每一帧
        for frame_file in frame_files:
            frame = cv2.imread(os.path.join(vis_dir, frame_file))
            out.write(frame)
        
        out.release()
        
        return jsonify({
            'success': True,
            'video_url': f"/api/videos/{video_id}_visualization.mp4"
        })
    
    except Exception as e:
        return jsonify({'error': f'生成可视化视频失败: {str(e)}'}), 500

@app.route('/api/videos/<filename>', methods=['GET'])
def get_visualization_video(filename):
    return send_from_directory(os.path.join(UPLOAD_FOLDER, 'videos'), filename)

if __name__ == '__main__':
    app.run(debug=True, host='0.0.0.0', port=5000)