1

.gitignore

@@ -1,2 +1,3 @@
 .idea/
 .vscode/
+0_roi_*.png

bim_cal.py

@@ -0,0 +1,296 @@
+import copy
+from typing import Optional
+from typing import Tuple
+import math
+
+
+def point_to_line_distance(px, py, x1, y1, x2, y2):
+    """
+    计算一个点，到另外两个点形成的直线上的垂直距离。（点到直线的距离公式）
+
+    返回距离值
+    """
+    # 分子部分
+    numerator = abs((y2 - y1) * px - (x2 - x1) * py + x2 * y1 - y2 * x1)
+    # 分母部分
+    denominator = math.hypot(y2 - y1, x2 - x1)
+    # 距离
+    return numerator / denominator
+
+
+def from_main_get_other_point_distance(main_ymj, other_ymj):
+    """
+    以main_ymj(主要预埋件)的四个边的中点。
+        高度中点连线为x轴
+        宽度中点连线为y轴
+    计算other_ymj(其他预埋件)的中心点到主要预埋件x轴直线和y轴直线的的垂直距离
+
+    返回x垂直距离,y垂直距离
+    """
+    # 其他预埋件的中心点坐标
+    px = other_ymj['x_center']
+    py = other_ymj['y_center']
+
+    # 计算其他预埋件中心点坐标到预埋件左右边中点坐标连线形成的直线的垂直距离
+    bottom_line_center_x = main_ymj['bottom_line_center_x']
+    bottom_line_center_y = main_ymj['bottom_line_center_y']
+    top_line_center_x = main_ymj['top_line_center_x']
+    top_line_center_y = main_ymj['top_line_center_y']
+
+    x_distance = point_to_line_distance(px, py, bottom_line_center_x, bottom_line_center_y, top_line_center_x,
+                                        top_line_center_y)
+
+    # 计算其他预埋件中心点坐标到预埋件左右边中点坐标连线形成的直线的垂直距离
+    left_line_center_x = main_ymj['left_line_center_x']
+    left_line_center_y = main_ymj['left_line_center_y']
+    right_line_center_x = main_ymj['right_line_center_x']
+    right_line_center_y = main_ymj['right_line_center_y']
+    y_distance = point_to_line_distance(px, py, left_line_center_x, left_line_center_y, right_line_center_x,
+                                        right_line_center_y)
+
+    return round(x_distance, 2), round(y_distance, 2)
+
+
+def from_main_get_other_point_distance_bim(main_ymj, other_ymj):
+    """
+    与from_main_get_other_point_distance函数类似
+    只不过计算的是bim图中，其他预埋件相对于主要预埋件的x和y的距离
+    因为bim是标准的平面图，所以可以直接使用主要预埋件的中心点作为坐标原点，直接计算其他预埋件相对于这个坐标原点的x,y,
+
+    返回x垂直距离,y垂直距离
+    """
+    main_x = main_ymj["x"]
+    main_x = float(main_x)
+
+    main_y = main_ymj["center"]
+    main_y = float(main_y) * 1000
+
+    other_x = other_ymj["x"]
+    other_x = float(other_x)
+
+    other_y = other_ymj["center"]
+    other_y = float(other_y)  * 1000
+
+    x_distance = abs(other_x - main_x)
+    y_distance = abs(other_y - main_y)
+
+    return round(x_distance, 2), round(y_distance, 2)
+
+
+def cal_deviation(main_ymj, other_ymj_list):
+    """
+    计算偏差表
+    bim和图片中其他预埋件中心点到主要预埋件的偏差值
+        中心x轴直线偏差值
+        中心y轴直线的的垂直距离偏差值
+        欧几里德距离偏差值
+
+
+    @:return
+        [
+            {
+            'ymj_code': 预埋件code
+            'x_deviation': x偏差
+            'y_deviation': y偏差
+            'euclidean_deviation': 欧几里德距离的偏差
+            'average_deviation' : 平均误差
+            'x_deviation_s': x偏差值，带正负符号
+            'y_deviation_s': y偏差值，带正负符号
+            }
+        ]
+    """
+    # 1.计算图片和bim中，其他预埋件中心点到主要预埋件中心x轴直线和中心y轴直线的的垂直距离
+    distance_data = []  # bim 和 图片中其他预埋件中心点到主要预埋件中心x轴直线和中心y轴直线的的垂直距离，欧几里德距离
+    for other_ymj in other_ymj_list:
+        # 计算图片中其他预埋件中心点到主要预埋件中心x轴直线和中心y轴直线的的垂直距离
+        x_distance_pic, y_distance_pic = from_main_get_other_point_distance(main_ymj, other_ymj)
+
+        pic_data = {
+            'x_distance': x_distance_pic,
+            'y_distance': y_distance_pic,
+            # 欧几里得距离值
+            'euclidean_distance': math.sqrt(x_distance_pic ** 2 + y_distance_pic ** 2),
+        }
+
+        # 计算bim中其他预埋件中心点到主要预埋件中心x轴直线和中心y轴直线的的垂直距离
+        x_distance_bim, y_distance_bim = from_main_get_other_point_distance_bim(main_ymj, other_ymj)
+        bim_data = {
+            'x_distance': x_distance_bim,
+            'y_distance': y_distance_bim,
+            # 欧几里得距离值
+            'euclidean_distance': math.sqrt(x_distance_bim ** 2 + y_distance_bim ** 2),
+        }
+
+        distance_data.append(
+            {
+                'code': other_ymj['code'],
+                'pic_data': pic_data,  # 图片中其他预埋件中心点到主要预埋件中心x轴直线和中心y轴直线的的垂直距离, 欧几里得距离值
+                'bim_data': bim_data,  # bim中其他预埋件中心点到主要预埋件中心x轴直线和中心y轴直线的的垂直距离, 欧几里得距离值
+            }
+        )
+
+    # 2.对比bim和图片中数据的偏差值
+    deviation_data = []  # 偏差值列表
+    for distance in distance_data:
+        pic_data = distance['pic_data']
+        bim_data = distance['bim_data']
+        ymj_code = distance['code']
+
+        # 计算x偏差
+        x_deviation = pic_data['x_distance'] - bim_data['x_distance']
+        # 计算y偏差
+        y_deviation = pic_data['y_distance'] - bim_data['y_distance']
+        # 计算欧几里得距离偏差
+        sqrt_xx_yy_deviation = pic_data['euclidean_distance'] - bim_data['euclidean_distance']
+        # 计算平均误差
+        average_deviation = max(abs(x_deviation), abs(y_deviation), abs(sqrt_xx_yy_deviation))
+
+        deviation_data.append({
+            'ymj_code': ymj_code,
+            'x_deviation': abs(x_deviation),
+            'x_deviation_s': x_deviation,
+            'y_deviation': abs(y_deviation),
+            'y_deviation_s': y_deviation,
+            'euclidean_deviation': abs(sqrt_xx_yy_deviation),
+            'average_deviation': average_deviation,
+        })
+
+    # 3.返回计算偏差值列表
+    return deviation_data
+
+
+
+def get_point_align_line(
+        px: float,
+        py: float,
+        x_axis: Optional[Tuple[float, float, float, float]],
+        y_axis: Optional[Tuple[float, float, float, float]],
+        x_move_distance: float,
+        y_move_distance: float
+) -> Tuple[float, float]:
+    """
+    计算一点沿着两点形成的直线方向，移动指定距离之后的坐标值。
+    注意方向为：
+        起点 x_axis_x1,x_axis_y1 ==> 终点 x_axis_x2,x_axis_y2
+        起点 y_axis_x1,y_axis_y1 ==> 终点 y_axis_x2,y_axis_y2
+    相当于 px, py 沿着 x_axis 和 y_axis 的方向各移动 x_move_distance 和 y_move_distance。
+
+    :return: 移动后的新坐标 (new_x, new_y)
+    """
+
+    new_x, new_y = px, py
+
+    # 如果 x 方向有位移
+    if x_move_distance != 0:
+        x1, y1, x2, y2 = x_axis
+        dx = x2 - x1
+        dy = y2 - y1
+        length = math.hypot(dx, dy)
+        if length != 0:
+            x_unit_vec = (dx / length, dy / length)
+            new_x += x_move_distance * x_unit_vec[0]
+            new_y += x_move_distance * x_unit_vec[1]
+
+    # 如果 y 方向有位移
+    if y_move_distance != 0:
+        x1, y1, x2, y2 = y_axis
+        dx = x2 - x1
+        dy = y2 - y1
+        length = math.hypot(dx, dy)
+        if length != 0:
+            y_unit_vec = (dx / length, dy / length)
+            new_x += y_move_distance * y_unit_vec[0]
+            new_y += y_move_distance * y_unit_vec[1]
+
+    return new_x, new_y
+
+def cal_deviation_table(all_ymj_data):
+    """
+    计算每个预埋件的平均偏差表，
+    :param all_ymj_data:
+    :return: [
+                {
+                    'ymj_code': 主要预埋件
+                    'average_deviation': 平均误差
+                }
+            ]
+    """
+    deviation_table = []  # 数值偏差表
+    for main_ymj in all_ymj_data:
+        other_ymj_list = copy.deepcopy(all_ymj_data)
+        other_ymj_list.remove(main_ymj)  # 移除主要预埋件
+
+        # 1.计算偏差数据
+        deviation_data = cal_deviation(main_ymj, other_ymj_list)
+        print(f"main_ymj_code={main_ymj['code']}")
+        print(f"deviation_data={deviation_data}")
+        # 2.计算所有其他预埋件误差的平均值
+        all_average_deviation = 0
+        for deviation in deviation_data:
+            # 一个预埋件的误差平均值
+            average_deviation = deviation['average_deviation']
+            # 误差 累加起来
+            all_average_deviation += average_deviation
+        # 除以预埋件数量
+        all_average_deviation = all_average_deviation / len(deviation_data)
+        deviation_table.append({
+            'ymj_code': main_ymj['code'],
+            'average_deviation': all_average_deviation,
+        })
+    print(f"📋deviation_table={deviation_table}")
+    return deviation_table
+
+
+
+
+def cal_with_manual_measurement(all_ymj_data):
+    """
+    根据人工测量出来的偏差值，计算预埋件偏差值
+
+    @:return
+        main_ymj_code : 使用了人工测量数据的预埋件
+        measure_deviation: 其他预埋件相当于人工测量出数据的实际偏差
+
+    """
+    main_ymj_code = None
+    measure_deviation = []
+    # 深度拷贝一份全部预埋件数据, 防止源数据被修改
+    all_ymj_data_copy = copy.deepcopy(all_ymj_data)
+    # 遍历查找哪个数据有人工测量的值
+    is_find = False
+    for idx, ymj in enumerate(all_ymj_data_copy):
+        # 如果有人工测量数据
+        if 'x_measure_bias' in ymj:
+            is_find = True
+            x_measure_bias = ymj['x_measure_bias']  # x偏差
+            y_measure_bias = ymj['y_measure_bias']  # y偏差
+            # 先获取其他预埋件相对于这个预埋件的偏差
+            other_ymj_list = all_ymj_data_copy
+            other_ymj_list.remove(ymj)
+            deviation_data = cal_deviation(ymj, other_ymj_list)
+            print(f"主预埋件code={ymj['code']}")
+            print(f"其他预埋件偏差={deviation_data}")
+            # 给其他预埋件的偏差值 加上人工测量的偏差 以获得其真实偏差
+            for deviation in deviation_data:
+                # 带符号的偏差偏差值
+                x_deviation_s = deviation['x_deviation_s']
+                y_deviation_s = deviation['y_deviation_s']
+                # 找到这个预埋件的源数据
+                for ymj_data in other_ymj_list:
+                    # code相同判断
+                    if ymj_data['code'] == deviation['ymj_code']:
+                        # 偏差值再加上人工测量的偏差 以获得其真实偏差
+                        measure_deviation.append({
+                            'ymj_code': deviation['ymj_code'],
+                            'x_deviation': x_deviation_s + x_measure_bias,
+                            'y_deviation': y_deviation_s + y_measure_bias,
+                        })
+                        break
+        if is_find:
+            # 只拿取第一个人工测量的预埋件的数据
+            main_ymj_code = ymj['code']
+            break
+
+    if not is_find:
+        print("❗️❗️没有任何一个预埋件存在人工测量的数据")
+    return main_ymj_code,measure_deviation

capture(1).py

@@ -0,0 +1,145 @@
+# -- coding: utf-8 --
+import time
+import os
+import paramiko
+from scp import SCPClient
+import threading
+from os import abort 
+
+
+HOST = "192.168.100.254"
+PORT = 22
+USERNAME = "root"
+PASSWORD = "ebaina"
+
+EXPTIME = 4500000
+CLOUDPOINTS = 30 * 10000
+
+
+def ssh_execute_command(host, port, username, password, command):
+    ret = False
+    # 创建SSH客户端
+    client = paramiko.SSHClient()
+    client.set_missing_host_key_policy(paramiko.AutoAddPolicy())
+    client.connect(host, port, username, password)
+ 
+    # 执行命令
+    stdin, stdout, stderr = client.exec_command(command)
+    print(stdout.read().decode())
+    print(stderr.read().decode())
+ 
+    # 关闭连接
+    client.close()
+    return True
+ 
+
+def scp_get_file(host, port, username, password, remote_path, local_path):
+    ret = False
+    # 创建SSH客户端
+    client = paramiko.Transport((host, port))
+    client.connect(username=username, password=password)
+ 
+    # 创建SCP客户端
+    scp_client = SCPClient(client)
+ 
+    # 拉取文件
+    try:
+        scp_client.get(remote_path, local_path)
+    except Exception as e:
+        print(e)
+        return False
+ 
+    # 关闭连接
+    scp_client.close()
+    client.close()
+
+    return True
+
+
+def cameraCap():
+    start_time = time.time() 
+    ret = ssh_execute_command(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        command = "cd /root/calibration_tools \n" +
+                  "./camera " + str(EXPTIME) + "\n"
+    )
+    end_time = time.time()  # 记录结束时间
+    execution_time = end_time - start_time  # 计算执行时间
+    print(f"ssh_execute_command 程序执行时间：{execution_time}秒")
+    time.sleep(1) # 等待操作系统保存图像完毕
+    ret = scp_get_file(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        remote_path = "/root/calibration_tools/output.png",
+        local_path = os.path.dirname(__file__) + "\\ssh_tmp\\output.png"
+    )
+    if not ret:
+        return False
+    ret = ssh_execute_command(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        command = "cd /root/calibration_tools \n " +
+                  "rm output.png \n" 
+    )
+    end_time = time.time()  # 记录结束时间
+    execution_time = end_time - start_time  # 计算执行时间
+    print(f"cameraCap 程序执行时间：{execution_time}秒")
+
+
+def lidarCap():
+    start_time = time.time() 
+    ret = ssh_execute_command(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        command = "cd /root/calibration_tools \n" +
+                  "./lidar_driver " + str(CLOUDPOINTS) + "\n"
+    )
+    ret = scp_get_file(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        remote_path = "/root/calibration_tools/output.ply",
+        local_path = os.path.dirname(__file__) + ".\\ssh_tmp\\output.ply"
+    )
+    ret = ssh_execute_command(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        command = "cd /root/calibration_tools \n " +
+                  "rm output.ply \n" 
+    )
+    end_time = time.time()  # 记录结束时间
+    execution_time = end_time - start_time  # 计算执行时间
+    print(f"lidarCap 程序执行时间：{execution_time}秒")
+
+
+def cap():
+
+    directory = os.path.dirname(__file__) + ".\\ssh_tmp\\"
+    files_and_directories = os.listdir(directory)
+    for file in files_and_directories:
+        file_path = os.path.join(directory, file)
+        if os.path.isfile(file_path):
+            os.unlink(file_path)
+
+    t_lidar = threading.Thread(target=lidarCap)
+    t_lidar.start()
+    cameraCap()
+
+    # t_camera.join()
+
+    t_lidar.join()
+
+if __name__ == '__main__':
+    cap()

capture_img_ply.py

@@ -17,8 +17,8 @@ CLOUD_COUNTS = [400000, 500000, 600000]
 DEFAULT_COUNT = 300000
 EXPS = [200 * 1000, 300 * 1000, 400 * 1000]
 DEFAULT_EXP = 100 * 1000
-PARENT_DIR_NAME = "test_2"
-SAVE_DIR_NAME = ""
+PARENT_DIR_NAME = "test_ymj"
+SAVE_DIR_NAME = "test_ymj_1"
 
 
 def ssh_execute_command(host, port, username, password, command):

capture_img_ply_txt._getFiel.py

@@ -15,7 +15,7 @@ USERNAME = "root"
 PASSWORD = "ebaina"
 
 # CLOUD_COUNTS = [600000,800000,1000000,1500000,2000000]
-CLOUD_COUNTS = [1500000]
+CLOUD_COUNTS = [300000]
 DEFAULT_COUNT = 1500000
 EXPS = [20*1000]
 # EXPS = [100 * 1000]

capture_img_ply_txt.py

@@ -15,12 +15,12 @@ USERNAME = "root"
 PASSWORD = "ebaina"
 
 # CLOUD_COUNTS = [600000,800000,1000000,1500000,2000000]
-CLOUD_COUNTS = [1500000]
-DEFAULT_COUNT = 1500000
-EXPS = [10*1000]
+CLOUD_COUNTS = []
+DEFAULT_COUNT = 300000
+EXPS = [4500000*1000]
 # EXPS = [100 * 1000]
-DEFAULT_EXP = int(8 * 1000)
-PARENT_DIR_NAME = "宁德核电站0327-9-长条形状2"
+DEFAULT_EXP = 0
+PARENT_DIR_NAME = "测试2"
 # PARENT_DIR_NAME = "test_1"
 SAVE_DIR_NAME = ""
 TEST_CLOUD = False

capture_new.py

@@ -0,0 +1,145 @@
+# -- coding: utf-8 --
+import time
+import os
+import paramiko
+from scp import SCPClient
+import threading
+from os import abort 
+
+
+HOST = "192.168.100.254"
+PORT = 22
+USERNAME = "root"
+PASSWORD = "ebaina"
+
+EXPTIME = 4500000
+CLOUDPOINTS = 30 * 10000
+
+
+def ssh_execute_command(host, port, username, password, command):
+    ret = False
+    # 创建SSH客户端
+    client = paramiko.SSHClient()
+    client.set_missing_host_key_policy(paramiko.AutoAddPolicy())
+    client.connect(host, port, username, password)
+ 
+    # 执行命令
+    stdin, stdout, stderr = client.exec_command(command)
+    print(stdout.read().decode())
+    print(stderr.read().decode())
+ 
+    # 关闭连接
+    client.close()
+    return True
+ 
+
+def scp_get_file(host, port, username, password, remote_path, local_path):
+    ret = False
+    # 创建SSH客户端
+    client = paramiko.Transport((host, port))
+    client.connect(username=username, password=password)
+ 
+    # 创建SCP客户端
+    scp_client = SCPClient(client)
+ 
+    # 拉取文件
+    try:
+        scp_client.get(remote_path, local_path)
+    except Exception as e:
+        print(e)
+        return False
+ 
+    # 关闭连接
+    scp_client.close()
+    client.close()
+
+    return True
+
+
+def cameraCap():
+    start_time = time.time() 
+    ret = ssh_execute_command(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        command = "cd /root/calibration_tools \n" +
+                  "./camera_driver " + str(EXPTIME) + "\n"
+    )
+    end_time = time.time()  # 记录结束时间
+    execution_time = end_time - start_time  # 计算执行时间
+    print(f"ssh_execute_command 程序执行时间：{execution_time}秒")
+    time.sleep(1) # 等待操作系统保存图像完毕
+    ret = scp_get_file(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        remote_path = "/root/calibration_tools/output.png",
+        local_path = os.path.dirname(__file__) + "\\ssh_tmp\\output.png"
+    )
+    if not ret:
+        return False
+    ret = ssh_execute_command(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        command = "cd /root/calibration_tools \n " +
+                  "rm output.png \n" 
+    )
+    end_time = time.time()  # 记录结束时间
+    execution_time = end_time - start_time  # 计算执行时间
+    print(f"cameraCap 程序执行时间：{execution_time}秒")
+
+
+def lidarCap():
+    start_time = time.time() 
+    ret = ssh_execute_command(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        command = "cd /root/calibration_tools \n" +
+                  "./lidar_driver " + str(CLOUDPOINTS) + "\n"
+    )
+    ret = scp_get_file(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        remote_path = "/root/calibration_tools/output.ply",
+        local_path = os.path.dirname(__file__) + ".\\ssh_tmp\\output.ply"
+    )
+    ret = ssh_execute_command(
+        host = HOST,
+        port = PORT,
+        username = USERNAME,
+        password = PASSWORD,
+        command = "cd /root/calibration_tools \n " +
+                  "rm output.ply \n" 
+    )
+    end_time = time.time()  # 记录结束时间
+    execution_time = end_time - start_time  # 计算执行时间
+    print(f"lidarCap 程序执行时间：{execution_time}秒")
+
+
+def cap():
+
+    directory = os.path.dirname(__file__) + ".\\ssh_tmp\\"
+    files_and_directories = os.listdir(directory)
+    for file in files_and_directories:
+        file_path = os.path.join(directory, file)
+        if os.path.isfile(file_path):
+            os.unlink(file_path)
+
+    t_lidar = threading.Thread(target=lidarCap)
+    t_lidar.start()
+    cameraCap()
+
+    # t_camera.join()
+
+    t_lidar.join()
+
+if __name__ == '__main__':
+    cap()

config.ini

@@ -12,7 +12,7 @@ p2 = -0.000424586368
 k3 = 0.371045956
 yolo_label = 0
 yolo_prob = 0.601
-yolo_modol_path = ./best_ep23_small_lab.om
+yolo_modol_path = ./ymj_03-04-enhance-712-lab-seaSlat.om
 save_image = true
 roi_y_offset = 20
 edge_min_line_len = 400
@@ -24,18 +24,18 @@ kk = 0.0
 r = 1.572895519320505
 p = -1.5570521101629837
 y = 0.024731696602587224
-tx = 0.109253
-ty = 0.0710213
-tz = 0.0175978
-dminx = -2.5
-dmaxx = 2.5
-dminy = -2.5
-dmaxy = 2.5
+tx = 0.110253
+ty = -0.0095978
+tz = 0.0910213
+dminx = -1.5
+dmaxx = 1.5
+dminy = -1.5
+dmaxy = 1.5
 dminz = 0.0
 dmaxz = 3.6
 kk = 0.0
-cloud_need_points_size = 600000
-save_cload = true
+cloud_need_points_size = 1500000
+save_cload = false
 
 [sys]
 fake = true
@@ -43,6 +43,12 @@ camera_cap_fake = true
 lidar_cap_fake = true
 npu_fake = true
 conners_detect_fake = true
-fake_image_fpath = ./result/hy_1/output.jpg
-fake_lidar_fpath = ./result/hy_1/output.ply
+fake_image_fpath = ./result2/test2/output.jpg
+fake_lidar_fpath = ./result2/test2/output.ply
+export_time = 24
+bim_angle_offset_max = 0.08
+bin_dist_offset_max = 0.03
+
+[test]
+apple = 234
 

conners.txt

@@ -0,0 +1,80 @@
+6683
+1675
+7716
+1675
+7716
+2654
+6683
+2654
+5799
+3730
+7895
+3730
+7895
+5796
+5799
+5796
+4469
+158
+5638
+158
+5638
+1310
+4469
+1310
+6594
+249
+7773
+249
+7773
+1384
+6594
+1384
+2031
+547
+3680
+547
+3680
+2151
+2031
+2151
+5244
+2321
+6266
+2321
+6266
+3314
+5244
+3314
+3505
+5140
+4863
+5140
+4863
+6515
+3505
+6515
+0
+18
+675
+18
+675
+1202
+0
+1202
+0
+1427
+463
+1427
+463
+2529
+0
+2529
+2132
+3680
+4847
+3680
+4847
+4909
+2132
+4909

detect.py

@@ -1,3 +1,4 @@
+import ctypes
 import os
 import shutil
 import sys
@@ -6,6 +7,8 @@ import time
 from ctypes import *
 import logging
 from logging import makeLogRecord
+
+import imgProcess
 from clog import logger
 import cv2
 import numpy as np
@@ -37,6 +40,46 @@ MSG_ONLY_RECORD = -1
 MSG_LOCATION_RECORD = 0
 MSG_DETECTION_RECORD = 1
 
+# bim_data = [
+#     {"code": "BFX1101VBPT001", "type": "248x248", "x": "325", "y": "250", "center": "1.269", "w": "248", "h": "248",
+#      "angle": "0°"},
+#     {"code": "BFX1101VBPT002", "type": "150x149", "x": "717", "y": "250", "center": "0.849", "w": "150", "h": "149",
+#      "angle": "0°"},
+#     {"code": "BFX1101VBPT003", "type": "300x248", "x": "778", "y": "250", "center": "0.317", "w": "300", "h": "248",
+#      "angle": "0°"},
+#     {"code": "BFX1101VBPT004", "type": "248x248", "x": "1668", "y": "250", "center": "0.296", "w": "248", "h": "248",
+#      "angle": "90°"},
+#     {"code": "BFX1101VBPT005", "type": "248x248", "x": "2454", "y": "250", "center": "1.307", "w": "248", "h": "248",
+#      "angle": "0°"}]
+
+bim_data = [
+    {"code": "BFX1101VBPT002", "type": "150x149", "x": "717", "y": "250", "center": "0.960", "w": "150", "h": "149",
+     "angle": "0°"},
+    {"code": "BFX1101VBPT003", "type": "300x248", "x": "778", "y": "250", "center": "0.317", "w": "300", "h": "248",
+     "angle": "0°"},
+    {"code": "BFX1101VBPT004", "type": "248x248", "x": "1668", "y": "250", "center": "0.296", "w": "248", "h": "248",
+     "angle": "90°"}]
+
+MAX_OBJ = 32
+
+
+class OjectInfo(ctypes.Structure):
+    _fields_ = [
+        ("id", ctypes.c_int),
+        ("name", ctypes.c_char * 256),
+        ("centerX", ctypes.c_float),
+        ("centerY", ctypes.c_float),
+        ("w", ctypes.c_float),
+        ("h", ctypes.c_float)
+    ]
+
+
+class OjectsInfo(ctypes.Structure):
+    _fields_ = [
+        ("count", ctypes.c_int),
+        ("objs", OjectInfo * MAX_OBJ)
+    ]
+
 
 # 获取消息
 def _get_msg_info(code) -> str:
@@ -71,6 +114,13 @@ class _SubRoiData_(Structure):
         ('mVc2D', POINTER(c_float)),
         ('mVc3D', POINTER(c_float)),
         ('H', c_float * 9),
+        # 新增部分
+        ('mInnerLineCenter', c_float * 8),  # 4 * 2
+        ('mInnerLineCenter3D', c_float * 12),  # 4 * 3
+        ('mInnerCenter', c_float * 2),  # 1 * 2
+        ('mInnerCenter3D', c_float * 3),  # 1 * 3
+        ('mEdgesLen', c_float * 4),  # 4
+        # 新增字段之后需同步在_copy_record函数中进行添加
     ]
 
 
@@ -106,22 +156,94 @@ def _copy_record(record):
         if i > 15:  # 超过16个ROI，返回
             break
         subRois = record.subRois[i]
+
         newSubRois = new_record.subRois[i]
+        newSubRois.isGood = subRois.isGood
+        newSubRois.mInfo = subRois.mInfo
         for j in range(8):
             newSubRois.mRoiPoints[j] = subRois.mRoiPoints[j]
             newSubRois.mInnerConners[j] = subRois.mInnerConners[j]
+            newSubRois.mInnerLineCenter[j] = subRois.mInnerLineCenter[j]
         for j in range(12):
             newSubRois.mInnerConners3D[j] = subRois.mInnerConners3D[j]
+            newSubRois.mInnerLineCenter3D[j] = subRois.mInnerLineCenter3D[j]
+        for j in range(4):
+            newSubRois.mEdgesLen[j] = subRois.mEdgesLen[j]
+        for j in range(2):
+            newSubRois.mInnerCenter[j] = subRois.mInnerCenter[j]
+        for j in range(3):
+            newSubRois.mInnerCenter3D[j] = subRois.mInnerCenter3D[j]
 
     return new_record
 
 
+# def init_callback():
+#     global data_callback, bStop
+#
+#     def _callback(data):
+#         print("_callback回调执行")
+#         global g_msg_cache, g_frame_cache, g_running_mode, bStop
+#         record = CallbackInfo(code=SYS_OK, errorInfo=b"", bGetData=False)
+#         frame = None
+#         ret = SYS_OK
+#         msg = None
+#
+#         # 获取数据
+#         data_type = CallbackInfo
+#         data_ptr = cast(data, POINTER(data_type))
+#         ckinfoCache = data_ptr.contents
+#         record = _copy_record(ckinfoCache)
+#         print(str(g_running_mode) + " : " + str(record.code) + " : " + bytes.decode(record.errorInfo))
+#         # 当定位模式下，接收到【定位完成】消息
+#         if g_running_mode == RUNNING_MODE_LOCATION and record.code == (6 + TO_CLIENT_NOTIFY_BASE):
+#             if len(g_frame_cache) == 0:
+#                 print("当前没有广角数据")
+#                 return
+#             frame = g_frame_cache.pop()  # 从图像缓存中去除图像
+#             g_frame_cache = []  # 清空缓存，避免无限增长
+#             msg = Msg(MSG_LOCATION_RECORD, _copy_record(record), frame)
+#         # 当检测模式下，接收到【拍照完成】消息
+#         elif record.code == (8 + TO_CLIENT_NOTIFY_BASE):
+#             bStop = True
+#             time.sleep(1)
+#         elif record.code == (12 + TO_CLIENT_NOTIFY_BASE):
+#             msg = Msg(MSG_DETECTION_RECORD, _copy_record(record), None)
+#             detectionResProcessing(msg,data_img_path,data_detect_res_save_path)
+#             bStop = True
+#             time.sleep(1)
+#         #     # frame = np.zeros((7000, 9344, 3), np.uint8)
+#         #     # # frame.zeros(9344, 7000, cv2.CV_8UC3)
+#         #     # # src_frame_ptr = c_char_p(frame.data.tobytes())
+#         #     # mv = memoryview(frame.data)
+#         #     # buffer_pointer = cast(mv.obj.ctypes.data, c_void_p)
+#         #     # ret = image_framework_sdk.LibapiGetHQImage(
+#         #     #     buffer_pointer, frame.shape[1], frame.shape[0], 3, c_char_p(b"PythonProcessing"))
+#         #     if ret != SYS_OK:
+#         #         print("ret != SYS_OK:" + str(ret))
+#         #         record = CallbackInfo(code=ret, errorInfo=_get_msg_info(ret), bGetData=False)
+#         #         msg = Msg(MSG_DETECTION_RECORD, record, None)
+#         #         g_msg_cache.append(msg)
+#         #         # emit msg
+#         #         return
+#         #     msg = Msg(MSG_DETECTION_RECORD, _copy_record(record), frame)
+#         # elif g_running_mode == RUNNING_MODE_LOCATION:
+#         #     msg = Msg(MSG_LOCATION_RECORD, _copy_record(record), None)
+#         # elif g_running_mode == RUNNING_MODE_DETECTION:
+#         #     msg = Msg(MSG_DETECTION_RECORD, _copy_record(record), None)
+#         # else:
+#         #     print("未知回调")
+#         #     return
+#
+#     data_callback = CallbackType(_callback)
+
 def init_callback():
-    global data_callback, bStop
+    global data_callback
 
     def _callback(data):
         print("_callback回调执行")
-        global g_msg_cache, g_frame_cache, g_running_mode, bStop
+
+        global g_msg_cache, g_frame_cache, g_running_mode
+
         record = CallbackInfo(code=SYS_OK, errorInfo=b"", bGetData=False)
         frame = None
         ret = SYS_OK
@@ -132,46 +254,55 @@ def init_callback():
         data_ptr = cast(data, POINTER(data_type))
         ckinfoCache = data_ptr.contents
         record = _copy_record(ckinfoCache)
-        print(str(g_running_mode) + " : " + str(record.code) + " : " + bytes.decode(record.errorInfo))
+        print(f"g_running_mode===={g_running_mode}")
+        print(f"record===={record.code}")
+        # print("callback ============ > " + str(record.code) + " mode " + str(g_running_mode))
+        # print(str(g_running_mode) + " : " + str(record.code) + " : " + bytes.decode(record.errorInfo))
         # 当定位模式下，接收到【定位完成】消息
         if g_running_mode == RUNNING_MODE_LOCATION and record.code == (6 + TO_CLIENT_NOTIFY_BASE):
             if len(g_frame_cache) == 0:
-                print("当前没有广角数据")
+                # self.logger.error("当前没有广角数据")
                 return
             frame = g_frame_cache.pop()  # 从图像缓存中去除图像
             g_frame_cache = []  # 清空缓存，避免无限增长
-            msg = Msg(MSG_LOCATION_RECORD, _copy_record(record), frame)
+            msg = Msg(MSG_LOCATION_RECORD, _copy_record(record), frame, None)
         # 当检测模式下，接收到【拍照完成】消息
-        elif record.code == (8 + TO_CLIENT_NOTIFY_BASE):
-            bStop = True
-            time.sleep(1)
-        elif record.code == (12 + TO_CLIENT_NOTIFY_BASE):
-            msg = Msg(MSG_DETECTION_RECORD, _copy_record(record), None)
-            detectionResProcessing(msg,data_img_path,data_detect_res_save_path)
-            bStop = True
-            time.sleep(1)
-        #     # frame = np.zeros((7000, 9344, 3), np.uint8)
-        #     # # frame.zeros(9344, 7000, cv2.CV_8UC3)
-        #     # # src_frame_ptr = c_char_p(frame.data.tobytes())
-        #     # mv = memoryview(frame.data)
-        #     # buffer_pointer = cast(mv.obj.ctypes.data, c_void_p)
-        #     # ret = image_framework_sdk.LibapiGetHQImage(
-        #     #     buffer_pointer, frame.shape[1], frame.shape[0], 3, c_char_p(b"PythonProcessing"))
-        #     if ret != SYS_OK:
-        #         print("ret != SYS_OK:" + str(ret))
-        #         record = CallbackInfo(code=ret, errorInfo=_get_msg_info(ret), bGetData=False)
-        #         msg = Msg(MSG_DETECTION_RECORD, record, None)
-        #         g_msg_cache.append(msg)
-        #         # emit msg
-        #         return
-        #     msg = Msg(MSG_DETECTION_RECORD, _copy_record(record), frame)
-        # elif g_running_mode == RUNNING_MODE_LOCATION:
-        #     msg = Msg(MSG_LOCATION_RECORD, _copy_record(record), None)
-        # elif g_running_mode == RUNNING_MODE_DETECTION:
-        #     msg = Msg(MSG_DETECTION_RECORD, _copy_record(record), None)
-        # else:
-        #     print("未知回调")
-        #     return
+        elif (g_running_mode == RUNNING_MODE_DETECTION and record.code == (2 + TO_CLIENT_NOTIFY_BASE)) or \
+                (g_running_mode == RUNNING_MODE_DETECTION and record.code == (8 + TO_CLIENT_NOTIFY_BASE)) or \
+                (g_running_mode == RUNNING_MODE_DETECTION and record.code == (12 + TO_CLIENT_NOTIFY_BASE)):
+            frame = np.zeros((7000, 9344, 3), np.uint8)
+            print("1")
+            # frame.zeros(9344, 7000, cv2.CV_8UC3)
+            # src_frame_ptr = c_char_p(frame.data.tobytes())
+            mv = memoryview(frame.data)
+            buffer_pointer = cast(mv.obj.ctypes.data, c_void_p)
+            print("12")
+            # ret = self.image_framework_sdk.LibapiGetHQImage(
+            #     buffer_pointer, frame.shape[1], frame.shape[0], 3, c_char_p(b"PythonProcessing"))
+            print("13")
+            if ret != SYS_OK:
+                print("ret != SYS_OK:" + str(ret))
+                record = CallbackInfo(code=ret, errorInfo=_get_msg_info(ret), bGetData=False)
+                msg = Msg(MSG_DETECTION_RECORD, record, None)
+                print("14")
+                g_msg_cache.append(msg)
+
+                # emit msg
+                return
+            msg = Msg(MSG_DETECTION_RECORD, _copy_record(record), frame)
+        elif g_running_mode == RUNNING_MODE_LOCATION:
+            # msg = Msg(MSG_LOCATION_RECORD, _copy_record(record), None, None)
+            print("")
+        elif g_running_mode == RUNNING_MODE_DETECTION:
+            # msg = Msg(MSG_DETECTION_RECORD, _copy_record(record), None, None)
+            print("")
+        else:
+            print("未知回调")
+            return
+
+        print("15")
+        g_msg_cache.append(msg)
+        # self.logger.info("image framework callback done.")
 
     data_callback = CallbackType(_callback)
 
@@ -188,10 +319,25 @@ def init_image_framework_sdk():
         print(f"Load Image framework sdk failed: {str(e)}")
 
 
+
+
 def adjust():
     global image_framework_sdk
     record = CallbackInfo(code=SYS_OK, errorInfo=b"", bGetData=False)
-    ret = image_framework_sdk.LibapiStartDetection()
+
+    # 创建一个 OjectsInfo 实例
+    info = OjectsInfo()
+
+    for idx, bim in enumerate(bim_data):
+        print(bim["center"])
+        info.objs[idx].id = idx
+        info.objs[idx].name = bim["code"].encode("utf-8")
+        info.objs[idx].centerX = float(bim["x"]) / 1000
+        info.objs[idx].centerY = float(bim["center"])
+        info.objs[idx].w = float(bim["w"]) / 1000
+        info.objs[idx].h = float(bim["h"]) / 1000
+
+    ret = image_framework_sdk.LibapiStartDetectionV2(ctypes.byref(info))
     if ret != SYS_OK:
         print(f"image_framework_sdk.LibapiStartDetection()：执行失败 :{ret}")
     return record
@@ -203,19 +349,23 @@ def check():
 
     record = CallbackInfo(code=SYS_OK, errorInfo=b"", bGetData=False)
     ret = image_framework_sdk.LibapiContinuetDetection("{}")
+    print("返回值:", ret)
     if ret != SYS_OK:
         print("image_framework_sdk.LibapiContinuetDetection：执行失败")
     return record
 
 
 def start():
+    global g_running_mode
     global bStop
+    g_running_mode = RUNNING_MODE_DETECTION
     # 开始检测
     logger.debug("开始adjust")
     adjust()
-    while not bStop:
-        logger.debug("adjust等待回调")
-        time.sleep(1)
+    # while not bStop:
+    #     logger.debug("adjust等待回调")
+    #     time.sleep(1)
+    time.sleep(5)
     logger.debug("开始check")
     check()
     bStop = False
@@ -224,7 +374,34 @@ def start():
         time.sleep(1)
 
 
+def _main_processing_thread():
+    global g_msg_cache, g_running_mode
+    record = CallbackInfo(code=SYS_OK, errorInfo=b"", bGetData=False)
+    times = 0
+    while True:
+        if len(g_msg_cache) <= 0:
+            times += 1
+            time.sleep(0.01)
+            # if times % 200 == 0:
+            # print("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+            continue
+
+        msg = g_msg_cache.pop()
+        # g_msg_cache = []
+        # print("===============> start" + str(msg.record.code))
+        if msg is None:
+            continue
+        msg = imgProcess.process(msg)
+        # print("===============> end" + str(msg.record.code))
+        # self.logger.info("msg ==> ui")
+        # self.signals.on_image_detect_result.emit(msg)
+
+
 if __name__ == '__main__':
+    process_thread = threading.Thread(target=_main_processing_thread)
+    process_thread.daemon = True
+    process_thread.start()
+
     # 加载驱动
     init_image_framework_sdk()
     if image_framework_sdk is None:
@@ -236,8 +413,7 @@ if __name__ == '__main__':
     if ret != 0:
         raise RuntimeError(f"设置回调函数始设置失败, 错误码: {ret}")
 
-
-    dataPath = "./result" # 数据路径
+    dataPath = "./result2"  # 数据路径
     data_img_path = None  # 检测的图片路径
     data_detect_res_save_path = None  # 检测结果的保存路径
     # 遍历目录

environment.yaml

@@ -1,4 +1,4 @@
-name: ymj_test
+name: ymj_data_collect
 channels:
   - defaults
   - https://repo.anaconda.com/pkgs/main
@@ -57,4 +57,3 @@ dependencies:
       - typing-extensions==4.12.2
       - tzdata==2025.1
       - zipp==3.21.0
-prefix: C:\Users\leon\miniconda3\envs\ymj_test

imgProcess.py

@@ -121,27 +121,158 @@ def doDetectionProcessing(msg):
         print("================== cv2.imwrite success =================")
 
         if record.code == 12 + TO_CLIENT_NOTIFY_BASE:
-            for j in range(record.roi_size):
-                roi = record.subRois[j]
-                if not roi.isGood: continue
-                ptsInnerConners = ptsInnerConnerslist[j]
-                ptsInnerConners3D = ptsInnerConners3Dlist[j]
-                if len(ptsInnerConners) == 4:
+            bim_result = []
+            for i in range(record.roi_size):
+                roi2 = record.subRois[i]
+                if not roi2.isGood: continue
+
+                mInfo = roi2.mInfo.decode('utf-8')
+                mInnerConners3D = list(roi2.mInnerConners3D)
+                mInnerLineCenter = list(roi2.mInnerLineCenter)
+                mInnerLineCenter3D = list(roi2.mInnerLineCenter3D)
+                mInnerCenter = list(roi2.mInnerCenter)
+                mInnerCenter3D = list(roi2.mInnerCenter3D)
+                mEdgesLen = list(roi2.mEdgesLen)
+                print(f"mEdgesLen={mEdgesLen}")
+                print(f"ℹ️ℹ️mInfo={mInfo}")
+                # 单位换算
+                for i in range(3):
+                    mInnerCenter3D[i] = mInnerCenter3D[i] * 1000
+                    mInnerCenter3D[i] = round(mInnerCenter3D[i], 2)
+                for i in range(12):
+                    # 米 单位 换算成毫米。# 留下小数
+                    mInnerConners3D[i] = mInnerConners3D[i] * 1000
+                    mInnerConners3D[i] = round(mInnerConners3D[i], 2)
+                    mInnerLineCenter3D[i] = mInnerLineCenter3D[i] * 1000
+                    mInnerLineCenter3D[i] = round(mInnerLineCenter3D[i], 2)
+                # mEdgesLen换算
                 for i in range(4):
-                        p3d = ptsInnerConners3D[i]
-                        p3d_next = ptsInnerConners3D[(i + 1) % 4]
-                        p2d = ptsInnerConners[i]
-                        p2d_next = ptsInnerConners[(i + 1) % 4]
-                        edge = np.sqrt(pow((p3d[0] - p3d_next[0]), 2)
-                                       + pow((p3d[1] - p3d_next[1]), 2)
-                                       + pow((p3d[2] - p3d_next[2]), 2))
-                        cv2.putText(im, str(edge)[:6], \
+                    # 米 单位 换算成毫米
+                    mEdgesLen[i] = mEdgesLen[i] * 1000
+                    # 留下小数
+                    mEdgesLen[i] = round(mEdgesLen[i], 2)
+
+                # 从bim数据数组中获取code等于mInfo的元素
+                # this_bim_data = None
+                # for bim_data_item in bim_data:
+                #     if str(bim_data_item["code"]) == mInfo:
+                #         this_bim_data = bim_data_item
+                #         break
+
+                # if this_bim_data is None:
+                #     print(f"🟡返回的值为mInfo={mInfo},无法在bim数据中找到code相同的")
+                #     continue
+                # else:
+                #     print(f"🟢{mInfo}预埋件匹配成功")
+
+                # 临时测试
+                bim_data = [
+                    {'code': 'BFX1101VBPT003', 'type': '300x248', 'x': '778', 'y': '250', 'center': '0.317', 'w': '300',
+                     'h': '248', 'angle': '0°', 'm_x': '0', 'm_y': '0', 'm_z': '0', 'base': '3'}]
+                this_bim_data = bim_data[0]
+                print(f"🟢{mInfo}预埋件匹配成功")
+
+                # bim的长度 和 计算出来的长度进行计算，取优的值
+                bim_data_width = float(this_bim_data["w"])
+                bim_data_height = float(this_bim_data["h"])
+                print(f"bim_data_width:{bim_data_width}")
+                print(f"bim_data_height:{bim_data_height}")
+                # 上右下左
+                cal_up_width = mEdgesLen[0]  # 上边宽度
+                cal_down_width = mEdgesLen[2]  # 下边宽度
+                cal_left_height = mEdgesLen[3]  # 左边高度
+                cal_right_height = mEdgesLen[1]  # 右边宽度
+                print("mEdgesLen:==", mEdgesLen)
+                # 取差值最小width 和 height的值
+                if abs(cal_up_width - bim_data_width) < abs(cal_down_width - bim_data_width):
+                    best_cal_width = cal_up_width
+                else:
+                    best_cal_width = cal_down_width
+
+                if abs(cal_left_height - bim_data_height) < abs(cal_right_height - bim_data_height):
+                    best_cal_height = cal_left_height
+                else:
+                    best_cal_height = cal_right_height
+
+                print("best_cal_width:==", best_cal_width)
+                print("best_cal_height:==", best_cal_height)
+                # 画到图上
+                cv2.putText(im, "w:" + str(best_cal_width), \
                             (int((ptsInnerConners[0][0] + ptsInnerConners[(0 + 1) % 4][0]) / 2), \
-                                     int((ptsInnerConners[0][1] + ptsInnerConners[(0 + 1) % 4][1]) / 2 + i * 20)), \
+                             int((ptsInnerConners[0][1] + ptsInnerConners[(0 + 1) % 4][1]) / 2 + 0 * 20)), \
                             cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (255, 0, 0), 2)
+
+                cv2.putText(im, "h:" + str(best_cal_height), \
+                            (int((ptsInnerConners[0][0] + ptsInnerConners[(0 + 1) % 4][0]) / 2), \
+                             int((ptsInnerConners[0][1] + ptsInnerConners[(0 + 1) % 4][1]) / 2 + 1 * 20)), \
+                            cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (255, 0, 0), 2)
+
+                # 预埋件基础数据
+                dict = this_bim_data
+                # 四个点的3d坐标
+                dict["x1"] = mInnerConners3D[0]
+                dict["y1"] = mInnerConners3D[1]
+                dict["z1"] = mInnerConners3D[2]
+                dict["x2"] = mInnerConners3D[3]
+                dict["y2"] = mInnerConners3D[4]
+                dict["z2"] = mInnerConners3D[5]
+                dict["x3"] = mInnerConners3D[6]
+                dict["y3"] = mInnerConners3D[7]
+                dict["z3"] = mInnerConners3D[8]
+                dict["x4"] = mInnerConners3D[9]
+                dict["y4"] = mInnerConners3D[10]
+                dict["z4"] = mInnerConners3D[11]
+                # 四条边中各自中点的3d坐标
+                dict["top_line_center_x"] = mInnerLineCenter3D[0]
+                dict["top_line_center_y"] = mInnerLineCenter3D[1]
+                dict["top_line_center_z"] = mInnerLineCenter3D[2]
+                dict["right_line_center_x"] = mInnerLineCenter3D[3]
+                dict["right_line_center_y"] = mInnerLineCenter3D[4]
+                dict["right_line_center_z"] = mInnerLineCenter3D[5]
+                dict["bottom_line_center_x"] = mInnerLineCenter3D[6]
+                dict["bottom_line_center_y"] = mInnerLineCenter3D[7]
+                dict["bottom_line_center_z"] = mInnerLineCenter3D[8]
+                dict["left_line_center_x"] = mInnerLineCenter3D[9]
+                dict["left_line_center_y"] = mInnerLineCenter3D[10]
+                dict["left_line_center_z"] = mInnerLineCenter3D[11]
+                # 计算出的宽度和高度
+                dict["best_cal_width"] = best_cal_width
+                dict["best_cal_height"] = best_cal_height
+                # 中心点3d坐标
+                dict["x_center"] = mInnerCenter3D[0]
+                dict["y_center"] = mInnerCenter3D[1]
+                dict["z_center"] = mInnerCenter3D[2]
+
+
+                dict["actual_value"] = ""
+                dict["status"] = "good"
+                dict["base"] = "5"
+                # dict["x_center"] =  mInnerConners3D[0]
+                # dict["y_center"] =  mInnerConners3D[1]
+                # dict["z_center"] =  mInnerConners3D[2]
+                bim_result.append(dict)
+            print(f"bim_result🟠={bim_result}")
+            # for j in range(record.roi_size):
+            #     roi = record.subRois[j]
+            #     if not roi.isGood: continue
+            #     ptsInnerConners = ptsInnerConnerslist[j]
+            #     ptsInnerConners3D = ptsInnerConners3Dlist[j]
+            #     if len(ptsInnerConners) == 4:
+            #         for i in range(4):
+            #             p3d = ptsInnerConners3D[i]
+            #             p3d_next = ptsInnerConners3D[(i + 1) % 4]
+            #             p2d = ptsInnerConners[i]
+            #             p2d_next = ptsInnerConners[(i + 1) % 4]
+            #             edge = np.sqrt(pow((p3d[0] - p3d_next[0]), 2)
+            #                            + pow((p3d[1] - p3d_next[1]), 2)
+            #                            + pow((p3d[2] - p3d_next[2]), 2))
+            #             cv2.putText(im, str(edge)[:6], \
+            #                         (int((ptsInnerConners[0][0] + ptsInnerConners[(0 + 1) % 4][0]) / 2), \
+            #                          int((ptsInnerConners[0][1] + ptsInnerConners[(0 + 1) % 4][1]) / 2 + i * 20)), \
+            #                         cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (255, 0, 0), 2)
         print("================== draw =================")
         msg.im2D = im
-        cv2.imshow("HQImage", im)
+        # cv2.imshow("HQImage", im)
         # cv2.waitKey(1)
     return msg
 
@@ -210,6 +341,11 @@ def detectionResProcessing(msg,data_img_path,data_detect_res_save_path):
     logger.success("================== cv2.imwrite result.jpg  success =================")
 
     if record.code == 12 + TO_CLIENT_NOTIFY_BASE:
+        for i in range(record.roi_size):
+            roi2 = record.subRois[i]
+            if not roi2.isGood: continue
+            mInfo = roi2.mInfo
+            print(f"mInfo====={mInfo}")
         for j in range(record.roi_size):
             roi = record.subRois[j]
             # if not roi.isGood: continue

result2/test_3/conners.txt

@@ -0,0 +1,24 @@
+2373
+866
+3068
+827
+3096
+1505
+2407
+1533
+2452
+3809
+3839
+3674
+3959
+4832
+2548
+4966
+6562
+3602
+7714
+3634
+7693
+4794
+6536
+4767

result2/test_3/roi_conners.txt

@@ -0,0 +1,26 @@
+2269
+754
+3219
+754
+3219
+1641
+2269
+1633
+2306
+3544
+4109
+3544
+4109
+5135
+2311
+5140
+6363
+3463
+7873
+3468
+7882
+4910
+6350
+4897
+
+

roi_conners.txt

@@ -1,64 +1,32 @@
-6911
-3175
-7935
-3175
-7935
-4142
-6911
-4142
-5578
-3808
-6623
-3808
-6623
-4761
-5578
-4761
-2677
-5089
-5214
-5089
-5214
-6179
-2677
-6179
-2613
-2198
-4142
-2198
-4142
-3607
-2613
-3607
-73
-1746
-1314
-1746
-1314
-2799
-73
-2799
-52
-3066
-1092
-3066
-1092
-3997
-52
-3997
-4815
-1805
-5951
-1805
-5951
-2849
-4815
-2849
-6774
-1829
-7896
-1829
-7896
-2869
-6774
-2869
+5708
+4469
+6930
+4469
+6930
+5660
+5708
+5660
+5
+1293
+1078
+1293
+1078
+2349
+5
+2349
+2551
+999
+3662
+999
+3662
+2054
+2551
+2054
+4666
+1307
+5807
+1307
+5807
+2406
+4666
+2406