Python Flask搭建yolov3目标检测系统详解流程_Python

Python Flask搭建yolov3目标检测系统详解流程

2022-02-25 00:26mind_programmonkey Python

YOLOv3没有太多的创新，主要是借鉴一些好的方案融合到YOLO里面。不过效果还是不错的，在保持速度优势的前提下，提升了预测精度，尤其是加强了对小物体的识别能力

【人工智能项目】Python Flask搭建yolov3目标检测系统：

Python Flask搭建yolov3目标检测系统详解流程

后端代码

from flask import Flask, request, jsonify
from PIL import Image
import numpy as np
import base64
import io
import os

from backend.tf_inference import load_model, inference

os.environ['CUDA_VISIBLE_DEVICES'] = '0'

sess, detection_graph = load_model()

app = Flask(__name__)

@app.route('/api/', methods=["POST"])
def main_interface():
  response = request.get_json()
  data_str = response['image']
  point = data_str.find(',')
  base64_str = data_str[point:]  # remove unused part like this: "data:image/jpeg;base64,"

  image = base64.b64decode(base64_str)       
  img = Image.open(io.BytesIO(image))

  if(img.mode!='RGB'):
      img = img.convert("RGB")
  
  # convert to numpy array.
  img_arr = np.array(img)

  # do object detection in inference function.
  results = inference(sess, detection_graph, img_arr, conf_thresh=0.7)
  print(results)

  return jsonify(results)

@app.after_request
def add_headers(response):
  response.headers.add('Access-Control-Allow-Origin', '*')
  response.headers.add('Access-Control-Allow-Headers', 'Content-Type,Authorization')
  return response


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