loop over frames from the video file stream
while True:
# 从文件中读取下一帧
(grabbed, frame) = vs.read()
# 如果帧没有被抓取,那么已经到了流的末尾
if not grabbed:
break
# 如果框架尺寸为空,则给他们赋值
if W is None or H is None:
(H, W) = frame.shape[:2]
# 从输入帧构造一个 blob,然后执行 YOLO 对象检测器的前向传递,得到边界框和相关概率
blob = cv2.dnn.blobFromImage(frame, 1 / 255.0, (416, 416),
swapRB=True, crop=False)
net.setInput(blob)
start = time.time()
layerOutputs = net.forward(outInfo)
end = time.time()
# 分别初始化检测到的边界框、置信度和类 ID 的列表
boxes = []
confidences = []
classIDs = []
# 循环输出
for output in layerOutputs:
# 遍历每个检测结果
for detection in output:
# 提取物体检测的类ID和置信度(即概率)
scores = detection[5:]
classID = np.argmax(scores)
confidence = scores[classID]
# 过滤精度低的结果
if confidence > confidence_t:
# 缩放边界框坐标,计算 YOLO 边界框的中心 (x, y) 坐标,然后是框的宽度和高度
box = detection[0:4] * np.array([W, H, W, H])
(centerX, centerY, width, height) = box.astype("int")
# 使用中心 (x, y) 坐标导出边界框的上角和左角
x = int(centerX - (width / 2))
y = int(centerY - (height / 2))
# 更新边界框坐标、[股指期货](https://www.gendan5.com/ff/sf.html)置信度和类 ID 列表
boxes.append([x, y, int(width), int(height)])
confidences.append(float(confidence))
classIDs.append(classID)
# 使用非极大值抑制来抑制弱的、重叠的边界框
idxs = cv2.dnn.NMSBoxes(boxes, confidences, confidence_t,
threshold)
# 确保至少存在一个检测
if len(idxs) > 0:
# 遍历保存的索引
for i in idxs.flatten():
# 在图像上绘制一个边界框矩形和标签
(x, y) = (boxes[i][0], boxes[i][1])
(w, h) = (boxes[i][2], boxes[i][3])
# 确保至少存在一个检测
color = [int(c) for c in COLORS[classIDs[i]]]
cv2.rectangle(frame, (x, y), (x + w, y + h), color, 2)
text = "{}: {:.4f}".format(LABELS[classIDs[i]],
confidences[i])
cv2.putText(frame, text, (x, y - 5),
cv2.FONT_HERSHEY_SIMPLEX, 0.5, color, 2)
cv2.imshow("Frame", frame)
key = cv2.waitKey(1) & 0xFF
# check if the video writer is None
if writer is None:
# initialize our video writer
fourcc = cv2.VideoWriter_fourcc(*'XVID')
writer = cv2.VideoWriter('output.avi', fourcc, 30, (int(frame.shape[1]), int(frame.shape[0])))
# some information on processing single frame
if total > 0:
elap = (end - start)
print("[INFO] single frame took {:.4f} seconds".format(elap))
print("[INFO] estimated total time to finish: {:.4f}".format(
elap * total))
# write the output frame to disk
writer.write(frame)
release the file pointers 【物体检测实战(使用 OpenCV 进行 YOLO 对象检测)】print("[INFO] cleaning up...")
writer.release()
vs.release()
推荐阅读
- 推荐系统论文进阶|CTR预估 论文精读(十一)--Deep Interest Evolution Network(DIEN)
- Python专栏|数据分析的常规流程
- Python|Win10下 Python开发环境搭建(PyCharm + Anaconda) && 环境变量配置 && 常用工具安装配置
- Python绘制小红花
- Pytorch学习|sklearn-SVM 模型保存、交叉验证与网格搜索
- OpenCV|OpenCV-Python实战(18)——深度学习简介与入门示例
- python|8. 文件系统——文件的删除、移动、复制过程以及链接文件
- 爬虫|若想拿下爬虫大单,怎能不会逆向爬虫,价值过万的逆向爬虫教程限时分享
- 分布式|《Python3网络爬虫开发实战(第二版)》内容介绍
- java|微软认真聆听了开源 .NET 开发社区的炮轰( 通过CLI 支持 Hot Reload 功能)