Upload 5 files

EDSR_x4.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd35ce3cae53ecee2d16045e08a932c3e7242d641bb65cb971d123e06904347f
+size 38573255

app.py

@@ -9,10 +9,14 @@ from sort import * #运动检测，采用sort算法
 import tempfile #创建输出临时文件夹
 import os
 from detectMotion import * #单独的运动检测
+import time
 
 
 #导入YoloV8模型,第一次使用会下载模型到当前文件夹当中
 model=YOLO("yolov8n.pt")
+global number_of_people_in_one_frame_list, flag
+number_of_people_in_one_frame_list = None
+flag = False
 
 #YoloV8标签数据，本次项目只使用了'person'
 classNames=['person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'traffic light',
@@ -31,6 +35,18 @@ classNames=['person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train'
 #运动检测算法参数
 tracker=Sort(max_age=20,min_hits=3,iou_threshold=0.3)
 
+def plot_number():
+    global number_of_people_in_one_frame_list, flag
+    if flag == False:
+        time.sleep(5)
+        flag = True
+    if number_of_people_in_one_frame_list == None:
+        raise gr.Error("请先选择需要处理的视频")
+    plt.close()
+    fig = plt.figure()
+    plt.plot(number_of_people_in_one_frame_list)
+    return fig
+
 # 彩色图像进行自适应直方图均衡化
 def hisEqulColor(img):
     fig_preprocessed = image_histogram(img)
@@ -48,23 +64,40 @@ def hisEqulColor(img):
     
     return img, fig_preprocessed, fig_postprocessed
 
-#添加高斯噪声，并使用中值滤波降噪
+#添加高斯噪声
 def AddGaussNoise(img,sigma):
     gauss=np.random.normal(0,sigma,img.shape)
     img=np.uint8(img + gauss)#将高斯噪声与原始图像叠加
-    img=cv2.medianBlur(img,5)
+    #img=cv2.medianBlur(img,5)
     return img
 
+#图像超分
+def AddISR(img):
+    sr = cv2.dnn_superres.DnnSuperResImpl_create()
+    sr.readModel("EDSR_x4.pb") 
+    sr.setModel("edsr", 4) # set the model by passing the value and the upsampling ratio
+    result = sr.upsample(img) # upscale the input image
+    return result
+
 #图像处理
-def processImg(img,sigma):
+def processImg(img,sigma,median_filter,ISR):
     img=cv2.cvtColor(img,cv2.COLOR_RGB2BGR)
     res1 = AddGaussNoise(img,sigma)
+    if median_filter == True:
+        #中值滤波降噪
+        img=cv2.medianBlur(img,5)
     res1, fig1, fig2 = hisEqulColor(res1)
+    if ISR == True:
+        res1 = AddISR(res1)
     res1=cv2.cvtColor(res1,cv2.COLOR_BGR2RGB)
     return res1, fig1, fig2
 
 #视频处理
 def processVideo(inputPath,codec):
+    global number_of_people_in_one_frame_list, flag
+    flag = False
+    number_of_people_in_one_frame = 0
+    number_of_people_in_one_frame_list = []
     number_of_people=0
     cap = cv2.VideoCapture(inputPath)#从inputPath读入视频
     fps = cap.get(cv2.CAP_PROP_FPS) #获取视频的帧率
@@ -87,6 +120,10 @@ def processVideo(inputPath,codec):
     elif codec == "wmv":
         fourcc = cv2.VideoWriter_fourcc('X','V','I','D')#视频编码：XVID,此编码不需要openh264-1.8.0-win64.dll
         video_save_path = os.path.join(outputPath,"output.wmv")#创建输出视频路径
+    elif codec == "mp4(h264)":
+        #h264:avc1, 在本地运行时可以在浏览器播放
+        fourcc = cv2.VideoWriter_fourcc('a','v','c','1')#视频编码：h264,只有h264格式的mp4文件才能在浏览器直接播放
+        video_save_path = os.path.join(outputPath,"output.mp4")#创建输出视频路径
 
     output_viedo.open(video_save_path , fourcc, fps, size, True)
     #对每一帧图片进行读取和处理
@@ -108,11 +145,15 @@ def processVideo(inputPath,codec):
                 cls=int(box.cls[0])#获取物体类别标签
                 #当标签为人，且可信度大于0.3的时候，将人标识出来
                 if cls==0 and conf > 0.3:
+                    number_of_people_in_one_frame += 1
                     cv2.rectangle(img,(x1,y1),(x2,y2),(255,0,255),3)
-                    print(conf)
+                    #print(conf)
                     #cvzone.putTextRect(img,f'{classNames[cls]}{conf}',(max(0,x1),max(30,y1)),scale=0.7,thickness=1)
                     currentArray=np.array([x1,y1,x2,y2,conf])
                     detections=np.vstack((detections,currentArray))#按行堆叠数据
+        number_of_people_in_one_frame_list.append(number_of_people_in_one_frame)
+        #print(type(number_of_people_in_one_frame_list))
+        number_of_people_in_one_frame = 0
         #运动检测
         resultsTracker=tracker.update(detections)
         for result in resultsTracker:
@@ -125,6 +166,7 @@ def processVideo(inputPath,codec):
         output_viedo.write(img)#将处理后的图像写入视频
     output_viedo.release()#释放
     cap.release()#释放
+    flag = False
     return video_save_path,video_save_path
 
 
@@ -149,17 +191,23 @@ with gr.Blocks() as demo:
         with gr.Row():
             with gr.Column():
                 text_inputPath = gr.Video()
-                codec = gr.Radio(["mp4","avi","mkv","wmv"], label="输出视频格式",
+                codec = gr.Radio(["mp4","avi","mkv","wmv",'mp4(h264)'], label="输出视频格式",
                                  value="mp4")
                 videoProcess_button = gr.Button("处理")
             with gr.Column():
                 text_output = gr.Video()
                 text_output_path = gr.Text(label="输出路径")
+        with gr.Row():
+            with gr.Column():
+                figure_number_output = gr.Plot(label="人流量")
+                #plotProcess_button = gr.Button("人流量显示")
     with gr.Tab("图像增强"):
         with gr.Row():
             with gr.Column():
                 image_input = gr.Image()
                 image_sigma = gr.Slider(0,40,label="高斯噪声sigma")
+                median_filter = gr.Checkbox(label="中值滤波")
+                Add_ISR = gr.Checkbox(label="图像超分")
             image_output = gr.Image()
             with gr.Column():
                 figure_pre_output = gr.Plot(label="处理前直方图")
@@ -180,10 +228,14 @@ with gr.Blocks() as demo:
 
     
 
-    videoProcess_button.click(processVideo, inputs=[text_inputPath,codec], outputs=[text_output,text_output_path])
-    image_button.click(processImg, inputs=[image_input,image_sigma], outputs=[image_output,figure_pre_output,figure_post_output])
-    motionProcess_button.click(motionDetection, inputs=[motion_inputPath], outputs=[motion_output_frame,motion_output_fmask,
-                                                                                frame_output_path,fmask_output_path])
+    videoProcess_button.click(processVideo, inputs=[text_inputPath,codec],
+                               outputs=[text_output,text_output_path])
+    image_button.click(processImg, inputs=[image_input,image_sigma,median_filter,Add_ISR], 
+                       outputs=[image_output,figure_pre_output,figure_post_output])
+    motionProcess_button.click(motionDetection, inputs=[motion_inputPath], 
+                               outputs=[motion_output_frame,motion_output_fmask,
+                                        frame_output_path,fmask_output_path])
+    videoProcess_button.click(plot_number,outputs=figure_number_output,every=1)
 
 demo.queue()#当有多个请求时，排队
 demo.launch()#生成内网链接，如需要公网链接，括号内输入share=True

requirements.txt

@@ -2,7 +2,9 @@ ultralytics
 opencv-python
 cvzone
 torch
+gradio
+opencv-contrib-python
 matplotlib
-filterpy==1.4.5
-scikit-image==0.17.2
-lap==0.4.0
+filterpy
+scikit-image
+lap