AI Label

DetectFineSortVideo.py

@@ -107,12 +107,12 @@ class DetFineVid():
                             tm2 = time.time()
                             if gp.DebugPrint_flag:
                                 print("main1 time is  ", (tm2 - tm1) * 1000)
-                            # cv2.imshow("AI-ECO", img_show)
+                            cv2.imshow("AI-ECO", img_show)
-                            # #cv2.imwrite("../data_MPKFNet/RTSP3/" + "{}.jpg".format(time.strftime('%Y%m%d%H%M%S', time.localtime(time.time()))),img_show)
+                            #cv2.imwrite("../data_MPKFNet/RTSP3/" + "{}.jpg".format(time.strftime('%Y%m%d%H%M%S', time.localtime(time.time()))),img_show)
-                            # if cv2.waitKey(1) & 0xFF == ord('r'):
+                            if cv2.waitKey(1) & 0xFF == ord('r'):
-                            #     break
+                                break
-                            save_path =os.path.join(self.SavePath,"{}.jpg".format(str(num_detection).zfill(6)))
+                            # save_path =os.path.join(self.SavePath,"{}.jpg".format(str(num_detection).zfill(6)))
-                            cv2.imwrite(save_path,img_show)
+                            # cv2.imwrite(save_path,img_show)
                             tm3 = time.time()
                             if gp.DebugPrint_flag:
                                 print("main2 time is  ", (tm3 - tm2) * 1000)

DetectionAILAbel.py

@@ -0,0 +1,30 @@
+import  os,argparse,torch
+from System_I import Detection_Y5
+from  tqdm import tqdm
+
+
+
+if __name__ == '__main__':
+    parser1 = argparse.ArgumentParser(description="AILabel")
+    parser1.add_argument("--RawPath", default='/data/DataManagement/Project/001_cuihu/0012_20220517_RTSP/imagesAILabel', type=str,
+                        help='RawPath.')
+    parser1.add_argument("--GPU", default="1", type=str,
+                        help='GPU.')
+    parser1.add_argument("--thresh", default=0.4, type=float,
+                        help='thresh.')
+    parser1.add_argument("--Weight", default="/data/Hatcher/data/Detection_weigth/BJCHKMCH-Bird-1-4/weights/best.pt", type=str,
+                        help='Weight.')
+    args = parser1.parse_args()
+    os.environ['CUDA_VISIBLE_DEVICES'] = args.GPU
+    devices = torch.device('cuda:{}'.format(str(int(args.GPU))) if torch.cuda.is_available() else 'cpu')
+    if not os.path.exists(args.RawPath.replace("imagesAILabel","NoResult")):
+        os.makedirs(args.RawPath.replace("imagesAILabel","NoResult"))
+    if not os.path.exists(args.RawPath.replace("imagesAILabel","YesResult")):
+        os.makedirs(args.RawPath.replace("imagesAILabel","YesResult"))
+    if not os.path.exists(args.RawPath.replace("imagesAILabel","LabelXML")):
+        os.makedirs(args.RawPath.replace("imagesAILabel","LabelXML"))
+    Detection = Detection_Y5(device=devices,root=" ", detect_thresh=args.thresh,DetWeight=args.Weight)
+    imagepaths = os.listdir(args.RawPath)
+    for imagepath in tqdm(imagepaths):
+        imagepath = os.path.join(args.RawPath,imagepath)
+        Detection.AIlabel(imagepath)

System_I.py

@@ -7,7 +7,7 @@ import json
 import random
 from configs.configs import model_head
 
-import os,sys
+import os,sys,shutil
 import torch
 from torch import nn
 from PIL import Image, ImageDraw, ImageFont
@@ -23,6 +23,7 @@ import numpy as np
 from fine.models.LoadModel import MainModel
 from fine.fineConfig import LoadConfig, load_data_transformers
 from torch.autograd import Variable
+from  creat_xml_fordetection import make_xml
 import threading
 from common import global_para as  gp
 import time,statistics
@@ -288,17 +289,14 @@ class Detection_S:
                 print("check final is ", (t5 - t4) * 1000)
         return flag,labels,boxes,scores
 class Detection_Y5:
-    def __init__(self,device,root):
+    def __init__(self,device,root,detect_thresh,DetWeight):
         self.device = device
         self.root = root
         self.img_size = 640
-        self.conf_thres = gp.detect_thresh
+        self.conf_thres = detect_thresh
         self.iou_thres = 0.4
         self.stride = 1
-        if gp.red_flag:
+        self.weights = DetWeight
-            self.weights = '../data_MPKFNet/test/2_r.pt'
-        else:
-            self.weights = gp.DetWeight
         # self.device = select_device(self.device)
         model = attempt_load(self.weights, map_location=self.device)
         model.to(self.device).eval()
@@ -320,6 +318,68 @@ class Detection_Y5:
             img = img.unsqueeze(0)
 
         return img0, img
+    def AIlabel(self,imagename):
+        with torch.no_grad():
+            image = cv2.imread(imagename)
+            image_or = image
+            img_detect = cv2.resize(image, (1024, 768))
+            t1 = time.time()
+            image_scale,image = self.preprocess(img_detect)
+            t2 = time.time()
+            if gp.DebugPrint_flag:
+                print("Detection data deal is ",(t2-t1)*1000)
+            pred = self.model(image, augment=False)[0]
+            pred = pred.float()
+            pred = non_max_suppression(pred, self.conf_thres, self.iou_thres)
+            t3 = time.time()
+            if gp.DebugPrint_flag:
+                print("Detection inference is ",(t3-t2)*1000)
+            boxes = []
+            scores = []
+            labels = []
+            for det in pred:
+                if det is not None and len(det):
+                    det[:, :4] = scale_coords(
+                        image.shape[2:], det[:, :4], image_scale.shape).round()
+
+                    boxes.append(det[:, :4])
+                    scores.append(det[:, 4])
+                    labels.append((det[:, -1]))
+                    boxes = boxes[0]
+                    scores = scores[0]
+                    labels = labels[0]
+            t5 = time.time()
+            if gp.DebugPrint_flag:
+                print("Detection get result is ", (t5 - t3) * 1000)
+            if len(labels) > 0:
+                ratio_x = (image_or.shape[1] / 1024.0)
+                ratio_y = (image_or.shape[0] / 768.0)
+
+
+                for i in range(len(boxes)):
+                    boxes[i][0]=boxes[i][0] * ratio_x
+                    boxes[i][2]=boxes[i][2] * ratio_x
+                    boxes[i][1] = boxes[i][1] * ratio_y
+                    boxes[i][3] = boxes[i][3] * ratio_y
+                    if boxes[i][0]<0:
+                        boxes[i][0] = 0
+                    if boxes[i][1]<0:
+                        boxes[i][1] = 0
+                    if boxes[i][2]<0:
+                        boxes[i][2] = 0
+                    if boxes[i][3]<0:
+                        boxes[i][3] = 0
+                # for i in range(len(labels)):
+                #     labels[i] = "bird"
+                t4 = time.time()
+                if gp.DebugPrint_flag:
+                    print("Detection final is ", (t4 - t3) * 1000)
+                xml_outputpath = imagename.replace("imagesAILabel","LabelXML").replace(".jpg",".xml")
+                make_xml(imagename,image_or.shape[0],image_or.shape[1],boxes,labels,scores,xml_outputpath)
+                shutil.copy(imagename, imagename.replace("imagesAILabel", "YesResult"))
+            else:
+                shutil.copy(imagename,imagename.replace("imagesAILabel","NoResult"))
+
     def inference(self,image,imagetime,index,fine_grained):
         with torch.no_grad():
             rootpath = os.path.join(self.root, imagetime)
@@ -1578,7 +1638,7 @@ class Density_mul (threading.Thread):
         gp.destory_image = True
         if gp.DebugPrint_flag:
             print(self.name + self.imagetime," is over")
-class Detecntion_mul (threading.Thread):
+class Detection_mul (threading.Thread):
     def __init__(self, threadID, name,imagetime,DetectionClass,Fine_grainedClass):
         threading.Thread.__init__(self)
         self.threadID = threadID

common.py

@@ -7,28 +7,28 @@
 import  os,torch, threading,argparse
 class global_para():
     #check()
-    parser1 = argparse.ArgumentParser(description="CUI")
+    # parser1 = argparse.ArgumentParser(description="CUI")
-    parser1.add_argument("video_num", default='/data1/Hatcher/model/project/data_MPKFNet/rawdata/竹林嘴-3月3日候鸟视频.mp4', type=str,
+    # parser1.add_argument("video_num", default='/data1/Hatcher/model/project/data_MPKFNet/rawdata/竹林嘴-3月3日候鸟视频.mp4', type=str,
-                        help='Video name.')
+    #                     help='Video name.')
-    parser1.add_argument("GPU", default='2', type=str,
+    # parser1.add_argument("GPU", default='2', type=str,
-                        help='GPU.')
+    #                     help='GPU.')
-    parser1.add_argument("--thresh", default=0.4, type=float,
+    # parser1.add_argument("--thresh", default=0.4, type=float,
-                        help='GPU.')
+    #                     help='GPU.')
-    parser1.add_argument("--red", default=0, type=int,
+    # parser1.add_argument("--red", default=0, type=int,
-                        help='red.')
+    #                     help='red.')
-    parser1.add_argument("--add", default=0, type=str,
+    # parser1.add_argument("--add", default=0, type=str,
-                        help='add.')
+    #                     help='add.')
-
+    #
-    args = parser1.parse_args()
+    # args = parser1.parse_args()
     #print(args)
-    root = "/home/gyd/output1"
+    # root = "/home/gyd/output1"
-    # os.environ['CUDA_VISIBLE_DEVICES'] = '0'
+    # # os.environ['CUDA_VISIBLE_DEVICES'] = '0'
-    os.environ['CUDA_VISIBLE_DEVICES'] = args.GPU
+    # os.environ['CUDA_VISIBLE_DEVICES'] = args.GPU
-    devices = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
+    # devices = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
-    camera_address = args.add
+    # camera_address = args.add
-    VideoPath = args.video_num
+    # VideoPath = args.video_num
-    red_flag = args.red
+    # red_flag = args.red
-    detect_thresh = args.thresh
+    # detect_thresh = args.thresh
 
     #fine_thresh2 = torch.tensor(fine_thresh2).cuda()
     birdIndex = 99999
@@ -51,7 +51,7 @@ class global_para():
 
     showThresh = 25
     destory_image = False
-    Camera_device = args.video_num  # ÉãÏñÍ·±àºÅ£¬0001
+    # Camera_device = args.video_num  # ÉãÏñÍ·±àºÅ£¬0001
     image_time = None  # µ±Ç°´¦ÀíµÄ30ÊÓƵµÄ±àºÅ
     ImgDealQueue = []  # 750Ö¡ÊÓƵ£¬»º´æ¿Õ¼ä
     img_show = None
@@ -80,28 +80,28 @@ class global_para():
 
     # DetWeight = "/data1/Hatcher/model/Detection/yolov5/runs/train/exp21/weights/best.pt"
     # FineWeight = "/data/Hatcher/data/weigth/areaproject/resnet50-JXPYH-Bird-3-4-20220429/_42911_PYH-3/best.pth"
-    DetWeight = "/data1/Hatcher/model/Detection/yolov5/runs/train/KunMingCuiHu-Fish2/weights/best.pt"
+    #DetWeight = "/data1/Hatcher/model/Detection/yolov5/runs/train/KunMingCuiHu-Fish2/weights/best.pt"
-    FineWeight = "/data/Hatcher/data/weigth/areaproject/resnet50-KMDC-Fish-5-1-20220505/_5510_DianChiFish-4/best.pth"
+    #FineWeight = "/data/Hatcher/data/weigth/areaproject/resnet50-KMDC-Fish-5-1-20220505/_5510_DianChiFish-4/best.pth"
-    # DetWeight = "/data1/Hatcher/model/Detection/yolov5/runs/train/BeijingCuiHu-Bird26/weights/best.pt"
+    DetWeight = "/data1/Hatcher/model/Detection/yolov5/runs/train/BeijingCuiHu-Bird26/weights/best.pt"
-    # FineWeight = "/data/Hatcher/data/weigth/areaproject/resnet50-cui90_1model0.7/_42111_CUI-94-1/best.pth"
+    FineWeight = "/data/Hatcher/data/weigth/areaproject/VIT-BJCH_Bird_94_DDP/_5914_CUI-94-1/best.pth"
     #BHG_chineseName = ["kong", '白鹤', '白头鹤','白枕鹤']
-    BHG_chineseName = ["kong","鲫鱼","麦穗鱼","鳑鲏","鲈鱼"]
+    #BHG_chineseName = ["kong","鲫鱼","麦穗鱼","鳑鲏","鲈鱼"]
 
-    # BHG_chineseName = ['Kong', '白额雁', '白骨顶', '苍鹭', '苍鹭', '白秋沙鸭', '白腰草鹬', '斑头雁', '斑嘴鸭', '苍鹭', '赤膀鸭', '赤颈鸭', '苍鹭', '大麻鳽',
+    BHG_chineseName = ['Kong', '白额雁', '白骨顶', '苍鹭', '苍鹭', '白秋沙鸭', '白腰草鹬', '斑头雁', '斑嘴鸭', '苍鹭', '赤膀鸭', '赤颈鸭', '苍鹭', '大麻鳽',
-    #                    '豆雁', '反嘴鹬', '凤头鸊鷉', '凤头麦鸡', '凤头潜鸭', '黑鹳', '黑水鸡', '红嘴鸥', '鸿雁', '灰鹤', '灰雁', '绿翅鸭', '绿头鸭', '琵嘴鸭',
+                       '豆雁', '反嘴鹬', '凤头鸊鷉', '凤头麦鸡', '凤头潜鸭', '黑鹳', '黑水鸡', '红嘴鸥', '鸿雁', '灰鹤', '灰雁', '绿翅鸭', '绿头鸭', '琵嘴鸭',
-    #                    '普通鸬鹚', '普通秋沙鸭', '翘鼻麻鸭', '鹊鸭', '扇尾沙锥', '小白额雁', '小天鹅', '针尾鸭', '苍鹭', '暗绿绣眼鸟', '八哥', '白腹鹞', '白鹡鸰',
+                       '普通鸬鹚', '普通秋沙鸭', '翘鼻麻鸭', '鹊鸭', '扇尾沙锥', '小白额雁', '小天鹅', '针尾鸭', '苍鹭', '暗绿绣眼鸟', '八哥', '白腹鹞', '白鹡鸰',
-    #                    '白眼潜鸭', '草鹭', '池鹭', '戴胜', '黑翅长脚鹬', '黑鸢', '红隼', '花脸鸭', '环颈雉', '灰头绿啄木鸟', '灰头麦鸡', '矶鹬', '栗苇鳽',
+                       '白眼潜鸭', '草鹭', '池鹭', '戴胜', '黑翅长脚鹬', '黑鸢', '红隼', '花脸鸭', '环颈雉', '灰头绿啄木鸟', '灰头麦鸡', '矶鹬', '栗苇鳽',
-    #                    '罗纹鸭', '牛背鹭', '普通翠鸟', '雀鹰', '喜鹊', '苍鹭', '鸳鸯', '棕背伯劳', '红胸秋沙鸭', '普通燕鸥', '小白额燕鸥', '丘鹬', '凤头蜂鹰',
+                       '罗纹鸭', '牛背鹭', '普通翠鸟', '雀鹰', '喜鹊', '苍鹭', '鸳鸯', '棕背伯劳', '红胸秋沙鸭', '普通燕鸥', '小白额燕鸥', '丘鹬', '凤头蜂鹰',
-    #                    '大天鹅', '大鸨', '小田鸡', '山斑鸠', '帆背潜鸭', '日本松雀鹰', '普通鵟', '松雀鹰', '毛脚鵟', '灰斑鸠', '白头鹞', '白尾鹞', '白眉鸭',
+                       '大天鹅', '大鸨', '小田鸡', '山斑鸠', '帆背潜鸭', '日本松雀鹰', '普通鵟', '松雀鹰', '毛脚鵟', '灰斑鸠', '白头鹞', '白尾鹞', '白眉鸭',
-    #                    '白鹈鹕', '红头潜鸭', '苍鹰', '蒙古银鸥', '西方秧鸡', '赤嘴潜鸭', '阔嘴鹬', '雕鸮', '鹌鹑', '鹗', '黄苇鳽', '黑颈鹤', '小䴙䴘', '赤麻鸭']
+                       '白鹈鹕', '红头潜鸭', '苍鹰', '蒙古银鸥', '西方秧鸡', '赤嘴潜鸭', '阔嘴鹬', '雕鸮', '鹌鹑', '鹗', '黄苇鳽', '黑颈鹤', '小䴙䴘', '赤麻鸭']
     # BHG_chineseName =["kong","绿头鸭","斑嘴鸭","灰雁","鸿雁","kong","小鷿鷈"]
     spiec_number_old = [0 for i in range(len(BHG_chineseName))]
-    FineClassNum = 5
+    FineClassNum = 94
     color_show = (255, 255, 255)
     color_box = (255, 0, 0)
-    NoneLabel = "鱼"#"鹤类" # "鱼"
+    NoneLabel = None#"鹤类" # "鱼"
-    fine_thresh1 = 6000
+    fine_thresh1 = 10000
-    fine_thresh2 = 6000
+    fine_thresh2 = 10000
     ShowAllNumFlag = True
     fontScale = 1
 """

creat_xml_fordetection.py

@@ -95,7 +95,7 @@ def make_xml(filename,size_x,size_y,boxes,labels,scores,savename):
         xmax = int(boxes[i][2])
         ymax = int(boxes[i][3])
         score = int(scores[i]*1000)
-        anno.add_pic_attr(str(labels[i]), xmin, ymin, xmax, ymax,score)
+        anno.add_pic_attr("bird", xmin, ymin, xmax, ymax,score)
     anno.savefile(savename)
 def make_xml_iccv(filename,size_x,size_y,boxes,savename):
     anno = GEN_Annotations(filename)

detection_negative_samples.py

@@ -3,16 +3,15 @@ import cv2
 import random
 from PIL import Image
 from lxml import etree
-path_unit= "/data1/zwq/model/data_MPKFNet/data/ConstructionSite/Hatcher_One"
+import argparse,random
-path_raw = "/data1/zwq/model/data_MPKFNet/data/ConstructionSite/NO"
+from tqdm import tqdm
-result_path = "/data1/zwq/model/data_MPKFNet/data/ConstructionSite/result"
 
 class GEN_Annotations:
     def __init__(self, filename):
         self.root = etree.Element("annotation")
 
         child1 = etree.SubElement(self.root, "folder")
-        child1.text = "VOC2007"
+        child1.text = "CDB"
 
         child2 = etree.SubElement(self.root, "filename")
         child2.text = filename
@@ -20,7 +19,7 @@ class GEN_Annotations:
         child3 = etree.SubElement(self.root, "source")
 
         child4 = etree.SubElement(child3, "annotation")
-        child4.text = "PASCAL VOC2007"
+        child4.text = "CDB"
         child5 = etree.SubElement(child3, "database")
         child5.text = "Unknown"
 
@@ -57,16 +56,21 @@ class GEN_Annotations:
         ymaxn.text = str(ymax)
 
 
-def negative_samples():
+def negative_samples(path_unit,path_raw):
     unit_list = os.listdir(path_unit)
+    random.shuffle(unit_list)
+    if len(unit_list) < 1000:
+        unit_list = unit_list
+    else:
+        unit_list = unit_list[:1000]
     unit_img_list = []
     for unit_img in unit_list:
         img1 = Image.open(os.path.join(path_unit,unit_img))
         unit_img_list.append(img1)
     raw_list = os.listdir(path_raw)
-    for raw_path  in raw_list:
+    for raw_path  in tqdm(raw_list):
-        print(raw_list)
+        # print(raw_list)
-        print("\n")
+        # print("\n")
         raw_img = Image.open(os.path.join(path_raw,raw_path))
         unit_num = random.randint(0,10)
         x_left = random.randint(0,raw_img.size[0]-unit_img_list[unit_num].size[0])
@@ -74,15 +78,33 @@ def negative_samples():
         x_right = x_left+unit_img_list[unit_num].size[0]
         y_right = y_left+unit_img_list[unit_num].size[1]
         raw_img.paste(unit_img_list[unit_num],(x_left,y_left,x_right,y_right))
-        raw_img.save(os.path.join(path_raw,raw_path).replace("/NO/","/result/Images/"))
+        raw_img.save(os.path.join(path_raw,raw_path).replace("/NOObject/","/result/Images/"))
         filename = raw_path
         anno = GEN_Annotations(filename)
         anno.set_size(raw_img.size[0], raw_img.size[1], 3)
         anno.add_pic_attr("person", x_left, y_left, x_right, y_right)
-        anno.savefile(os.path.join(path_raw,raw_path).replace("/NO/","/result/Ground_truth/").replace(".jpg",".xml"))
+        anno.savefile(os.path.join(path_raw,raw_path).replace("/NOObject/","/result/Ground_truth/").replace(".jpg",".xml"))
 
 
 
 if __name__ == '__main__':
+    parser1 = argparse.ArgumentParser(description="CUI")
+    parser1.add_argument("path_unit", default='/data/DataManagement/Project/001_cuihu/CAS_CuihuBird/CMB_Cuihubird/FinalData/data',
+                         type=str,
+                         help='path to object unit')
+    parser1.add_argument("path_raw", default='/data/DataManagement/Project/001_cuihu/0009_20220507_RTSP/CDB_CuihuNobird/NOObject', type=str,
+                         help='path to image to deal with')
 
-    negative_samples()
+
+    args = parser1.parse_args()
+    path_unit = args.path_unit
+    path_raw = args.path_raw
+    save_path1 = path_raw.replace("NOObject","result/Images/")
+    print(save_path1)
+    if not os.path.exists(save_path1):
+        os.makedirs(save_path1)
+    save_path2 = path_raw.replace("NOObject","result/Ground_truth/")
+    print(save_path2)
+    if not os.path.exists(save_path2):
+        os.makedirs(save_path2)
+    negative_samples(path_unit,path_raw)