基于PaddleOCR的数字显示器字符识别

项目介绍

在电力工业的100多年历史中，“智能电表”仍处于部署和使用的初级阶段。 由于电表是电力公司的前端“收银机”，必须十分精确。

尽管对于电表有大量的要求，性能规范和法规，并且电力公司试图确保电表精度，但实际上，一旦电表设计经过认证、制造和安装，大多数电表的精度的确认仅停留于生产刚结束时的出厂测试阶段。 特定电表在退役前的现场性能只能通过统计样本测试来估计。

此外，窃电是电力公司收入损失的主要根源。 虽然通常认为面临这一问题的主要是发展中经济体，但在许多发达地区也日益严重。 例如，英国天然气和电力市场办公室（Ofgem）在2013年7月发布一份题为“阻止窃电–诊断”的报告，其中估计英国每年发生的窃电价值超过2亿英镑，电力公司还需另外投资2500万英镑防窃电、修理或更换被篡改的设备。

在这里插入图片描述

了解整个电表的精度很关键

智能电表相对于电子和机械电表的主要优势是连通性。 联网智能电表可远程报告用电量、实施断电管理、收集使用时间数据并防止某些类型的窃电。 不过，有没有可能对电表本身关键的测量功能执行更精密的诊断？

其他具有关键任务功能的行业，例如汽车和工业，对诊断要求引入了“功能安全性”概念，其本质是检查设备以确定其在使用前、使用中和使用后是否正常运行。 电力电表行业的此类功能之一是现场使用寿命期间的电表精度。

目前电表执行现场样本测试，仅依靠现场电表内部元器件仍保持在校准范围内来估计精度，但此方法存在风险。 现场精度监控很重要，因为精度受传感器影响，传感器则暴露于高电流、电压事件和恶劣的环境。 因此，诊断必须包括监控整个电表，包括传感器和所有电子元件。

电表精度检查通常需要人为干预、断开现场连接和采用专用设备，因此需投入大量成本或实施复杂的拆解才能在现场完成。

针对以上这些分析，使用PaddleOCR可以处理并解决部分场景下的电表读数问题。

# 导入依赖库
import os
from tqdm import tqdm
import cv2
import csv
from PIL import Image
import matplotlib.pyplot as plt
import warnings
warnings.filterwarnings("ignore")

!pip install paddleocr==2.4.*

数据处理

# 解压测试数据
!unzip -oq /home/aistudio/data/data169810/test.zip  -d /home/aistudio/work/dataset

在进行电表读数之前，先进行数据可视化

数据可视化

# 可视化绘图
def imshow_image(img_path):
    img = Image.open(img_path)
    plt.figure("test_img", figsize=(5,5))
    plt.imshow(img)
    plt.show()

# 数据可视化
img_path= "/home/aistudio/work/dataset/test/133102_steerPoint5_preset1255_20220917221726_v.jpeg"
imshow_image(img_path)

在这里插入图片描述

问题分析
要处理电表中的数据，可以分为步骤，拆解为以下问题：

  • 1.感兴趣区域定位问题
  • 2.OCR读数问题

针对问题1,经过实验与探索，也找到两种方案：

方案1是，直接利用PaddleOCR默认自带的检测器，筛选掉其他无效的框体和信息，剩下的就是有用的。(未经过训练的，直接使用预训练模型)

方案2是，通过Opencv图像处理的方法，根据电表字符区域特征进行相应的轮廓提取和颜色筛选，从而保证其得到有效的定位。

方案3是，收集场景下的大量电表字符识别数据，制作数据集并进行标记，分别训练其定位和识别模型。

考虑到时间成本和人工成本问题，这里优先选择前两种方案。
下面是使用Opencv来进行ROI区域定位的方法。

定位OCR区域

import cv2
import numpy as np
import os

def find_biggest_contour(image):
    """获取最大轮廓"""
    image = image.copy()
    contours, hierarchy = cv2.findContours(image, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
    contour_sizes = [(cv2.contourArea(contour), contour) for contour in contours]
    biggest_contour = max(contour_sizes, key=lambda x: x[0])[1]
    return biggest_contour

def get_find_display(input_path, lower=(0, 0, 0), higher=(255, 255, 255), output_path='./'):
    """查找ROI轮廓"""
    img = cv2.imread(input_path)
    # print('input:', input_path)
    filename = input_path.split('/')[-1]
    f_name = filename.split('.')[0]
    # print('filename:', filename, 'f_name:', f_name)

    global img_crop
    lowHue = lower[0]
    lowSat = lower[1]
    lowVal = lower[2]
    highHue = higher[0]
    highSat = higher[1]
    highVal = higher[2]

    # 可选择不同的模糊方法
    frameBGR = cv2.GaussianBlur(img, (7, 7), 0)

    # 转换为HSV颜色空间
    hsv = cv2.cvtColor(frameBGR, cv2.COLOR_BGR2HSV)

    # 定义HSV值颜色范围
    colorLow = np.array([lowHue, lowSat, lowVal])
    colorHigh = np.array([highHue, highSat, highVal])
    mask = cv2.inRange(hsv, colorLow, colorHigh)

    kernal = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (7, 7))
    mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernal)
    mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernal)

    biggest_contour = find_biggest_contour(mask)
    # cv2.drawContours(img, biggest_contour, -1, (0, 255, 0), 2)
    print('cnt_len:', len(biggest_contour))

    # 将遮罩放在原始图像的上方。
    result_img = cv2.bitwise_and(img, img, mask=mask)

    if biggest_contour is not None:
        x, y, w, h = cv2.boundingRect(biggest_contour)
        print(x, y, w, h)
        img_crop = img[y:y + h, x:x + w]
        
        print('wpath:', output_path+filename)
        save_path = output_path + filename
        if not os.path.exists(output_path):
            os.mkdir(output_path)
        cv2.imwrite(save_path, img_crop)
    else:
        img_crop = img

    return result_img, img_crop

output_path = '/home/aistudio/work/roi/'
img_roi = '/home/aistudio/work/dataset/test/133102_steerPoint5_preset1255_20220917221726_v.jpeg'
lower = (0, 80, 0)
higher = (255, 255, 255)
result_img, img_crop = get_find_display(img_roi, lower, higher, output_path)

cnt_len: 573
67 33 483 208
wpath: /home/aistudio/work/roi/133102_steerPoint5_preset1255_20220917221726_v.jpeg

下图为本地运行结果，仅供参考。

在这里插入图片描述

在jupyter中运行查看下保存的结果

img_roi_path= "/home/aistudio/work/roi/133102_steerPoint5_preset1255_20220917221726_v.jpeg"
imshow_image(img_roi_path)

在这里插入图片描述

得到上图这样的结果后，可以直接将图片丢到PaddleOCR自带方法去识别。

from paddleocr import PaddleOCR

ocr = PaddleOCR()
def rec_display_roi(img_roi): 
    ocr = PaddleOCR() 
    result = ocr.ocr(img_roi, det=False) 
    return result[0][0], result[0][1]
rec_display_roi(img_roi_path)

方案2，直接使用PaddleOCR将所有可能是OCR的对象进行检测和识别。
再从中筛选要的结果。

初探OCR识别读数

from paddleocr import PaddleOCR, draw_ocr

# Paddleocr目前支持的多语言语种可以通过修改lang参数进行切换
# 例如`ch`, `en`, `fr`, `german`, `korean`, `japan`
ocr = PaddleOCR(use_angle_cls=True, lang="ch")  # need to run only once to download and load model into memory
img_path = '/home/aistudio/work/dataset/test/133102_steerPoint5_preset1255_20220917221726_v.jpeg'
save_path = '/home/aistudio/work/dst/result.jpg'

result = ocr.ocr(img_path, cls=True)
for line in result:
    print(line)

# 显示结果
from PIL import Image

image = Image.open(img_path).convert('RGB')
boxes = [line[0] for line in result]
txts = [line[1][0] for line in result]
scores = [line[1][1] for line in result]
im_show = draw_ocr(image, boxes, txts, scores, font_path='/home/aistudio/work/font/simfang.ttf')
im_show = Image.fromarray(im_show)
im_show.save(save_path)
imshow_image(save_path)


[2022/09/22 14:46:03] ppocr DEBUG: Namespace(alpha=1.0, benchmark=False, beta=1.0, cls_batch_num=6, cls_image_shape='3, 48, 192', cls_model_dir='/home/aistudio/.paddleocr/whl/cls/ch_ppocr_mobile_v2.0_cls_infer', cls_thresh=0.9, cpu_threads=10, crop_res_save_dir='./output', det=True, det_algorithm='DB', det_db_box_thresh=0.6, det_db_score_mode='fast', det_db_thresh=0.3, det_db_unclip_ratio=1.5, det_east_cover_thresh=0.1, det_east_nms_thresh=0.2, det_east_score_thresh=0.8, det_fce_box_type='poly', det_limit_side_len=960, det_limit_type='max', det_model_dir='/home/aistudio/.paddleocr/whl/det/ch/ch_PP-OCRv2_det_infer', det_pse_box_thresh=0.85, det_pse_box_type='quad', det_pse_min_area=16, det_pse_scale=1, det_pse_thresh=0, det_sast_nms_thresh=0.2, det_sast_polygon=False, det_sast_score_thresh=0.5, draw_img_save_dir='./inference_results', drop_score=0.5, e2e_algorithm='PGNet', e2e_char_dict_path='./ppocr/utils/ic15_dict.txt', e2e_limit_side_len=768, e2e_limit_type='max', e2e_model_dir=None, e2e_pgnet_mode='fast', e2e_pgnet_score_thresh=0.5, e2e_pgnet_valid_set='totaltext', enable_mkldnn=False, fourier_degree=5, gpu_mem=500, help='==SUPPRESS==', image_dir=None, ir_optim=True, label_list=['0', '180'], label_map_path='./vqa/labels/labels_ser.txt', lang='ch', layout_label_map=None, layout_path_model='lp://PubLayNet/ppyolov2_r50vd_dcn_365e_publaynet/config', max_batch_size=10, max_seq_length=512, max_text_length=25, min_subgraph_size=15, mode='structure', model_name_or_path=None, ocr_version='PP-OCRv2', output='./output', precision='fp32', process_id=0, rec=True, rec_algorithm='CRNN', rec_batch_num=6, rec_char_dict_path='/home/aistudio/.data/webide/pip/lib/python3.7/site-packages/paddleocr/ppocr/utils/ppocr_keys_v1.txt', rec_image_shape='3, 32, 320', rec_model_dir='/home/aistudio/.paddleocr/whl/rec/ch/ch_PP-OCRv2_rec_infer', save_crop_res=False, save_log_path='./log_output/', scales=[8, 16, 32], show_log=True, structure_version='STRUCTURE', table_char_dict_path=None, table_char_type='en', table_max_len=488, table_model_dir=None, total_process_num=1, type='ocr', use_angle_cls=True, use_dilation=False, use_gpu=False, use_mp=False, use_onnx=False, use_pdserving=False, use_space_char=True, use_tensorrt=False, vis_font_path='./doc/fonts/simfang.ttf', warmup=False)
[2022/09/22 14:46:05] ppocr DEBUG: dt_boxes num : 4, elapse : 0.8229670524597168
[2022/09/22 14:46:05] ppocr DEBUG: cls num  : 4, elapse : 0.19943690299987793
[2022/09/22 14:46:06] ppocr DEBUG: rec_res num  : 4, elapse : 0.5032625198364258
[[[186.0, 87.0], [451.0, 103.0], [446.0, 186.0], [182.0, 170.0]], ('3942', 0.80900913)]
[[[64.0, 222.0], [151.0, 225.0], [150.0, 254.0], [63.0, 250.0]], ('SEPM', 0.68886095)]
[[[469.0, 233.0], [506.0, 233.0], [506.0, 282.0], [469.0, 282.0]], ('V', 0.9785347)]
[[[71.0, 252.0], [150.0, 255.0], [149.0, 276.0], [71.0, 273.0]], ('南自仪表', 0.9353922)]

在这里插入图片描述

算法优化

def write_to_csv(log_path, filename='', result=0.00, score=0, mode_head=True):
    file = open(log_path, 'a+', encoding='utf-8', newline='')
    csv_writer = csv.writer(file)
    if mode_head == True:
        csv_writer.writerow([f'filename', f'result', f'score'])
    else:
        csv_writer.writerow([filename, result, score])
    file.close()

def get_bbox_area(box):
    """计算bbox的面积"""
    bbox_area = (max(box[2]) - max(box[0])) * (max(box[3]) - max(box[1]))
    return bbox_area

def quadArea(nodes):
    """计算多边形的面积"""
    # 基于向量积计算不规则多边形的面积, 坐标点需要按顺序（逆时针或顺时针）选取
    i_count = len(nodes)
    area_temp = 0
    for i in range(i_count):
        area_temp += nodes[i][0] * nodes[(i+1) % i_count][1] - nodes[(i+1) % i_count][0] * nodes[i][1]
    return abs(area_temp)

def bboxes_choose(boxes,txts,scores):
    """获取最大框体"""
    area_list = []
    for i in range(0, len(boxes)):
        bx = boxes[i]
        # area = get_bbox_area(bx)
        area = quadArea(bx)
        # print('bx:', bx, 'area:',area)
        area_list.append(area)

    if len(area_list) == 0:
        index = 0
    else:
        index = area_list.index(max(area_list))

    if len(boxes) == 0:
        boxes = []
    else:
        boxes = [boxes[index]]
        txts = [txts[index]]
        scores = [scores[index]]
    return boxes, txts, scores

def ocr_roi_det(img_path, font,save_path='./work/save/'):
    """OCR识别"""
    result = ocr.ocr(img_path, cls=True)
    # for line in result:
        # print(line)

    # 显示结果
    from PIL import Image
    image = Image.open(img_path).convert('RGB')

    fileslist = img_path.split('/')
    fname = fileslist[-1].split('.')[0]

    # [[[151.0, 53.0], [277.0, 53.0], [277.0, 111.0], [151.0, 111.0]], ('00.2', 0.9423570036888123)]
    boxes = [line[0] for line in result]
    txts = [line[1][0] for line in result]
    scores = [line[1][1] for line in result]
    boxes, txts, scores = bboxes_choose(boxes, txts, scores)
    # bs = nms(boxes, scores)
    # print('bs:', bs)

    im_show = draw_ocr(image, boxes, txts, scores, font_path=font)
    im_show = Image.fromarray(im_show)
    if not os.path.exists(save_path):
        os.mkdir(save_path)
    im_show.save(save_path + fname + '_result.jpg')
    
    return txts[0], scores[0] 

def all_test_det(path, log_path, font, save_path):
    """执行识别算法，并记录结果到csv"""
    count = 0
    img_list = []
    img_ans_dic = {}
    for filepath, dirnames, filenames in os.walk(path):  # 在多级目录下找文件
        for filename in filenames:
            file_path = filepath + filename
            # print('file_path:', file_path)
            img_list.append(file_path)

    global score
    write_to_csv(log_path)
    for i in tqdm(range(0, len(img_list)-1)):
        img_roi = img_list[i]
        # result, score = rec_display_roi(img_roi)
        fileslist = img_roi.split('/')
        fname = fileslist[-1].split('.')[0]
        result, score = ocr_roi_det(img_roi, font,save_path)
        print('result:', result, 'score:', score)

    
        if result != '':
            img_ans_dic[fname] = score
            count += 1
        else:
            score = -1
            img_ans_dic[fname] = score
            continue
        write_to_csv(log_path, fname, result, score,False)

    print('count:', count)
    print('dict_len:', len(img_ans_dic))
    print('ans_dict:', img_ans_dic)

if __name__ == '__main__':
    # Paddleocr目前支持的多语言语种可以通过修改lang参数进行切换
    # 例如`ch`, `en`, `fr`, `german`, `korean`, `japan`
    ocr = PaddleOCR(use_angle_cls=False, lang="en")  # need to run only once to download and load model into memory

    font = '/home/aistudio/work/font/simfang.ttf'
    # ocr_roi_det(img_path, font)

    log_path = '/home/aistudio/work/log/result.csv'
    save_path = '/home/aistudio/work/save_result/'
    test_path = '/home/aistudio/work/dataset/test/'
    all_test_det(test_path, log_path, font ,save_path)


[2022/09/22 15:12:37] ppocr DEBUG: Namespace(alpha=1.0, benchmark=False, beta=1.0, cls_batch_num=6, cls_image_shape='3, 48, 192', cls_model_dir='/home/aistudio/.paddleocr/whl/cls/ch_ppocr_mobile_v2.0_cls_infer', cls_thresh=0.9, cpu_threads=10, crop_res_save_dir='./output', det=True, det_algorithm='DB', det_db_box_thresh=0.6, det_db_score_mode='fast', det_db_thresh=0.3, det_db_unclip_ratio=1.5, det_east_cover_thresh=0.1, det_east_nms_thresh=0.2, det_east_score_thresh=0.8, det_fce_box_type='poly', det_limit_side_len=960, det_limit_type='max', det_model_dir='/home/aistudio/.paddleocr/whl/det/en/en_ppocr_mobile_v2.0_det_infer', det_pse_box_thresh=0.85, det_pse_box_type='quad', det_pse_min_area=16, det_pse_scale=1, det_pse_thresh=0, det_sast_nms_thresh=0.2, det_sast_polygon=False, det_sast_score_thresh=0.5, draw_img_save_dir='./inference_results', drop_score=0.5, e2e_algorithm='PGNet', e2e_char_dict_path='./ppocr/utils/ic15_dict.txt', e2e_limit_side_len=768, e2e_limit_type='max', e2e_model_dir=None, e2e_pgnet_mode='fast', e2e_pgnet_score_thresh=0.5, e2e_pgnet_valid_set='totaltext', enable_mkldnn=False, fourier_degree=5, gpu_mem=500, help='==SUPPRESS==', image_dir=None, ir_optim=True, label_list=['0', '180'], label_map_path='./vqa/labels/labels_ser.txt', lang='en', layout_label_map=None, layout_path_model='lp://PubLayNet/ppyolov2_r50vd_dcn_365e_publaynet/config', max_batch_size=10, max_seq_length=512, max_text_length=25, min_subgraph_size=15, mode='structure', model_name_or_path=None, ocr_version='PP-OCRv2', output='./output', precision='fp32', process_id=0, rec=True, rec_algorithm='CRNN', rec_batch_num=6, rec_char_dict_path='/home/aistudio/.data/webide/pip/lib/python3.7/site-packages/paddleocr/ppocr/utils/en_dict.txt', rec_image_shape='3, 32, 320', rec_model_dir='/home/aistudio/.paddleocr/whl/rec/en/en_number_mobile_v2.0_rec_infer', save_crop_res=False, save_log_path='./log_output/', scales=[8, 16, 32], show_log=True, structure_version='STRUCTURE', table_char_dict_path=None, table_char_type='en', table_max_len=488, table_model_dir=None, total_process_num=1, type='ocr', use_angle_cls=False, use_dilation=False, use_gpu=False, use_mp=False, use_onnx=False, use_pdserving=False, use_space_char=True, use_tensorrt=False, vis_font_path='./doc/fonts/simfang.ttf', warmup=False)









  0%|          | 0/14 [00:00<?, ?it/s][A[A[A[A[A[A[A

[2022/09/22 15:12:38] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:39] ppocr DEBUG: dt_boxes num : 15, elapse : 1.394296646118164
[2022/09/22 15:12:41] ppocr DEBUG: rec_res num  : 15, elapse : 1.3982703685760498









  7%|▋         | 1/14 [00:03<00:40,  3.09s/it][A[A[A[A[A[A[A

result: 482 score: 0.97532207
[2022/09/22 15:12:41] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:42] ppocr DEBUG: dt_boxes num : 16, elapse : 1.394862413406372
[2022/09/22 15:12:44] ppocr DEBUG: rec_res num  : 16, elapse : 1.2905313968658447









 14%|█▍        | 2/14 [00:05<00:36,  3.03s/it][A[A[A[A[A[A[A

result: 480 score: 0.97698164
[2022/09/22 15:12:44] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:45] ppocr DEBUG: dt_boxes num : 18, elapse : 1.4989678859710693
[2022/09/22 15:12:47] ppocr DEBUG: rec_res num  : 18, elapse : 1.3994457721710205









 21%|██▏       | 3/14 [00:09<00:33,  3.05s/it][A[A[A[A[A[A[A

result: 475 score: 0.9991196
[2022/09/22 15:12:47] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:48] ppocr DEBUG: dt_boxes num : 4, elapse : 0.9958655834197998
[2022/09/22 15:12:48] ppocr DEBUG: rec_res num  : 4, elapse : 0.29001855850219727









 29%|██▊       | 4/14 [00:10<00:25,  2.56s/it][A[A[A[A[A[A[A

result: 15.6 score: 0.99385107
[2022/09/22 15:12:49] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:50] ppocr DEBUG: dt_boxes num : 18, elapse : 1.2016396522521973
[2022/09/22 15:12:51] ppocr DEBUG: rec_res num  : 18, elapse : 1.4024155139923096









 36%|███▌      | 5/14 [00:13<00:23,  2.66s/it][A[A[A[A[A[A[A

result: 555 score: 0.9947159
[2022/09/22 15:12:51] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:52] ppocr DEBUG: dt_boxes num : 4, elapse : 0.8874359130859375
[2022/09/22 15:12:53] ppocr DEBUG: rec_res num  : 4, elapse : 0.29741787910461426









 43%|████▎     | 6/14 [00:14<00:18,  2.28s/it][A[A[A[A[A[A[A

result: 394.2 score: 0.9848536
[2022/09/22 15:12:53] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:53] ppocr DEBUG: dt_boxes num : 4, elapse : 0.3028900623321533
[2022/09/22 15:12:53] ppocr DEBUG: rec_res num  : 4, elapse : 0.2898275852203369









 50%|█████     | 7/14 [00:15<00:12,  1.81s/it][A[A[A[A[A[A[A

result: 260 score: 0.99791867
[2022/09/22 15:12:53] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:54] ppocr DEBUG: dt_boxes num : 4, elapse : 0.30339860916137695
[2022/09/22 15:12:54] ppocr DEBUG: rec_res num  : 4, elapse : 0.29735636711120605









 57%|█████▋    | 8/14 [00:16<00:09,  1.51s/it][A[A[A[A[A[A[A

result: 10.2 score: 0.9172772
[2022/09/22 15:12:54] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:56] ppocr DEBUG: dt_boxes num : 3, elapse : 1.5035269260406494
[2022/09/22 15:12:56] ppocr DEBUG: rec_res num  : 3, elapse : 0.19791388511657715











result: 002 score: 0.9978652


 64%|██████▍   | 9/14 [00:18<00:08,  1.65s/it][A[A[A[A[A[A[A

[2022/09/22 15:12:56] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:57] ppocr DEBUG: dt_boxes num : 2, elapse : 0.4907827377319336
[2022/09/22 15:12:57] ppocr DEBUG: rec_res num  : 2, elapse : 0.19646453857421875









 71%|███████▏  | 10/14 [00:19<00:05,  1.40s/it][A[A[A[A[A[A[A

result: 00.0 score: 0.8969887
[2022/09/22 15:12:57] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:58] ppocr DEBUG: dt_boxes num : 4, elapse : 1.2975883483886719
[2022/09/22 15:12:59] ppocr DEBUG: rec_res num  : 4, elapse : 0.2966597080230713









 79%|███████▊  | 11/14 [00:20<00:04,  1.52s/it][A[A[A[A[A[A[A

result: 894 score: 0.99809813
[2022/09/22 15:12:59] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:12:59] ppocr DEBUG: dt_boxes num : 4, elapse : 0.39928150177001953
[2022/09/22 15:13:00] ppocr DEBUG: rec_res num  : 4, elapse : 0.20896530151367188









 86%|████████▌ | 12/14 [00:21<00:02,  1.33s/it][A[A[A[A[A[A[A

result: 001 score: 0.9889035
[2022/09/22 15:13:00] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:13:01] ppocr DEBUG: dt_boxes num : 4, elapse : 0.9879066944122314
[2022/09/22 15:13:01] ppocr DEBUG: rec_res num  : 4, elapse : 0.29622650146484375









 93%|█████████▎| 13/14 [00:23<00:01,  1.35s/it][A[A[A[A[A[A[A

result: 393.9 score: 0.941982
[2022/09/22 15:13:01] ppocr WARNING: Since the angle classifier is not initialized, the angle classifier will not be uesd during the forward process
[2022/09/22 15:13:03] ppocr DEBUG: dt_boxes num : 16, elapse : 1.3959689140319824
[2022/09/22 15:13:04] ppocr DEBUG: rec_res num  : 16, elapse : 1.198702335357666









100%|██████████| 14/14 [00:26<00:00,  1.82s/it]

result: 48.1. score: 0.94031876


[A[A[A[A[A[A[A






[A[A[A[A[A[A[A

count: 14
dict_len: 14
ans_dict: {'249553_steerPoint7_preset1257_20220917131810_v': 0.97532207, '58867_steerPoint7_preset1257_20220917161848_v': 0.97698164, '268997_steerPoint7_preset1257_20220917194818_v': 0.9991196, '72635_steerPoint12_preset1294_20220919123447_v': 0.99385107, '217734_steerPoint6_preset1256_20220917121753_v': 0.9947159, '133102_steerPoint5_preset1255_20220917221726_v': 0.9848536, '265968_steerPoint3_preset1241_20220918001329_v': 0.99791867, '333810_steerPoint3_preset1241_20220918170603_v': 0.9172772, '241047_steerPoint1_preset1229_20220919115945_v': 0.9978652, '274368_steerPoint3_preset1244_20220917184302_v': 0.8969887, '50375_steerPoint1_preset1232_20220918013638_v': 0.99809813, '71585_steerPoint9_preset1263_20220918191815_v': 0.9889035, '187206_steerPoint8_preset1262_20220917115119_v': 0.941982, '133444_steerPoint7_preset1257_20220917151834_v': 0.94031876}

结果分析

rs_img = '/home/aistudio/work/save_result/133102_steerPoint5_preset1255_20220917221726_v_result.jpg'
imshow_image(rs_img)

在这里插入图片描述

import pandas as pd

csv_path = '/home/aistudio/work/log/result.csv'
df_data = pd.read_csv(csv_path)
print(df_data.info())
print(df_data.describe())

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 14 entries, 0 to 13
Data columns (total 3 columns):
 #   Column    Non-Null Count  Dtype  
---  ------    --------------  -----  
 0   filename  14 non-null     object 
 1   result    14 non-null     object 
 2   score     14 non-null     float64
dtypes: float64(1), object(2)
memory usage: 464.0+ bytes
None
           score
count  14.000000
mean    0.971728
std     0.033699
min     0.896989
25%     0.950317
50%     0.986879
75%     0.997078
max     0.999120

sample_for_plotting = df_data[['filename','result','score']]
data_head = sample_for_plotting.head()
data_head
filenameresultscore
0249553_steerPoint7_preset1257_20220917131810_v4820.975322
158867_steerPoint7_preset1257_20220917161848_v4800.976982
2268997_steerPoint7_preset1257_20220917194818_v4750.999120
372635_steerPoint12_preset1294_20220919123447_v15.60.993851
4217734_steerPoint6_preset1256_20220917121753_v5550.994716

总结

遇到一个算法识别问题的时候，要学会利用掌握的知识和经验，将问题进行拆解与转化。

并且要考虑到实际应用场景下的问题，有些步骤和条件需不需要？是否只能用深度学习的方法？

能不能用其他方式来进行某个方面的辅助，这些都是值得去思考和研究的问题。

此文章为搬运
原项目链接

# 人工智能
Logo
百度飞桨AI Studio社区

学大模型，用大模型上飞桨星河社区！每天8点V100G算力免费领！免费领取ERNIE 4.0 100w Token >>>

更多推荐

cover

Python调用文心一言的API

avatar 百度飞桨星河社区
cover

文心一言API调用（python）

avatar 百度飞桨星河社区

[PaddleSpeech 原神] 音色克隆之胡桃

[PaddleSpeech] 音色克隆之原神角色 ，项目代码已完整放在aistudio,地址如下：快去体验吧，可以直接运行。

avatar 百度飞桨星河社区