pdf扫描件表格识别不准确，如何优化？

[PureWaterCatt]

问题描述 / Problem Description

版面分析+表格识别，包括尝试单独截下表格来识别，都无法识别到最后一行的数据。
在版面分析的时候，表格也是被完整的table（label）的矩形方框包裹着。

运行环境 / Runtime Environment

Paddle：develop
PaddleOCR：develop
OS: ubuntu 20.04
GCC version: (Ubuntu/Linaro 8.4.0-3ubuntu2) 8.4.0
Clang version: N/A
CMake version: version 3.27.7
Libc version: glibc 2.31
Python version: 3.10.14

复现代码 / Reproduction Code

python predict_table.py --image_dir=../../output/screenshot-20240528-164403.png --det_model_dir=../inference/ch_PP-OCRv4_det_infer --rec_model_dir=../inference/ch_PP-OCRv4_rec_infer --rec_char_dict_path=../../ppocr/utils/ppocr_keys_v1.txt --table_model_dir=../inference/ch_ppstructure_mobile_v2.0_SLANet_infer --table_char_dict_path=../../ppocr/utils/dict/table_structure_dict_ch.txt --output=../../output/table

完整报错 / Complete Error Message

可能解决方案 / Possible solutions

附件 / Appendix

test.zip

[GreatV]

@PureWaterCatt 能提供原始图片吗

[PureWaterCatt]

@PureWaterCatt 能提供原始图片吗
@GreatV

[GreatV]

[2024/05/28 20:11:42] ppocr DEBUG: Namespace(help='==SUPPRESS==', use_gpu=False, use_xpu=False, use_npu=False, use_mlu=False, ir_optim=True, use_tensorrt=False, min_subgraph_size=15, precision='fp32', gpu_mem=500, gpu_id=0, image_dir=None, page_num=0, det_algorithm='DB', det_model_dir='/Users/wangxin/.paddleocr/whl/det/ch/ch_PP-OCRv4_det_infer', det_limit_side_len=960, det_limit_type='max', det_box_type='quad', det_db_thresh=0.3, det_db_box_thresh=0.6, det_db_unclip_ratio=1.5, max_batch_size=10, use_dilation=False, det_db_score_mode='fast', det_east_score_thresh=0.8, det_east_cover_thresh=0.1, det_east_nms_thresh=0.2, det_sast_score_thresh=0.5, det_sast_nms_thresh=0.2, det_pse_thresh=0, det_pse_box_thresh=0.85, det_pse_min_area=16, det_pse_scale=1, scales=[8, 16, 32], alpha=1.0, beta=1.0, fourier_degree=5, rec_algorithm='SVTR_LCNet', rec_model_dir='/Users/wangxin/.paddleocr/whl/rec/ch/ch_PP-OCRv4_rec_infer', rec_image_inverse=True, rec_image_shape='3, 48, 320', rec_batch_num=6, max_text_length=25, rec_char_dict_path='/Users/wangxin/repos/PaddleOCR/ppocr/utils/ppocr_keys_v1.txt', use_space_char=True, vis_font_path='./doc/fonts/simfang.ttf', drop_score=0.5, e2e_algorithm='PGNet', e2e_model_dir=None, e2e_limit_side_len=768, e2e_limit_type='max', e2e_pgnet_score_thresh=0.5, e2e_char_dict_path='./ppocr/utils/ic15_dict.txt', e2e_pgnet_valid_set='totaltext', e2e_pgnet_mode='fast', use_angle_cls=False, cls_model_dir=None, cls_image_shape='3, 48, 192', label_list=['0', '180'], cls_batch_num=6, cls_thresh=0.9, enable_mkldnn=False, cpu_threads=10, use_pdserving=False, warmup=False, sr_model_dir=None, sr_image_shape='3, 32, 128', sr_batch_num=1, draw_img_save_dir='./inference_results', save_crop_res=False, crop_res_save_dir='./output', use_mp=False, total_process_num=1, process_id=0, benchmark=False, save_log_path='./log_output/', show_log=True, use_onnx=False, return_word_box=False, output='./output', table_max_len=488, table_algorithm='TableAttn', table_model_dir='/Users/wangxin/.paddleocr/whl/table/ch_ppstructure_mobile_v2.0_SLANet_infer', merge_no_span_structure=True, table_char_dict_path='/Users/wangxin/repos/PaddleOCR/ppocr/utils/dict/table_structure_dict_ch.txt', layout_model_dir='/Users/wangxin/.paddleocr/whl/layout/picodet_lcnet_x1_0_fgd_layout_cdla_infer', layout_dict_path='/Users/wangxin/repos/PaddleOCR/ppocr/utils/dict/layout_dict/layout_cdla_dict.txt', layout_score_threshold=0.5, layout_nms_threshold=0.5, kie_algorithm='LayoutXLM', ser_model_dir=None, re_model_dir=None, use_visual_backbone=True, ser_dict_path='../train_data/XFUND/class_list_xfun.txt', ocr_order_method=None, mode='structure', image_orientation=True, layout=True, table=True, ocr=True, recovery=False, use_pdf2docx_api=False, invert=False, binarize=False, alphacolor=(255, 255, 255), lang='ch', det=True, rec=True, type='ocr', savefile=False, ocr_version='PP-OCRv4', structure_version='PP-StructureV2')
[2024/05/28 20:11:42] ppcls WARNING: The current running environment does not support the use of GPU. CPU has been used instead.
[2024/05/28 20:11:44] ppocr DEBUG: dt_boxes num : 43, elapsed : 0.09825301170349121
[2024/05/28 20:11:47] ppocr DEBUG: rec_res num  : 43, elapsed : 3.3741910457611084
[2024/05/28 20:11:48] ppocr DEBUG: dt_boxes num : 43, elapse : 0.09876203536987305
[2024/05/28 20:11:51] ppocr DEBUG: rec_res num  : 43, elapse : 3.1829168796539307
{'type': 'table', 'bbox': [0, 0, 458, 361], 'res': {'cell_bbox': [[5.933736324310303, 3.846806526184082, 151.30047607421875, 4.098109722137451, 151.35910034179688, 90.56180572509766, 5.676283836364746, 90.94571685791016], [147.81179809570312, 3.6569559574127197, 225.90658569335938, 3.77463698387146, 227.0962371826172, 79.76953125, 147.03846740722656, 80.49063110351562], [233.22488403320312, 3.6309869289398193, 284.1729736328125, 3.738816022872925, 285.87506103515625, 77.31891632080078, 234.9944610595703, 77.89393615722656], [279.1354675292969, 4.3987555503845215, 321.6705017089844, 4.548232555389404, 323.4425048828125, 78.41556549072266, 280.9509582519531, 78.96294403076172], [335.737060546875, 4.784805774688721, 362.3387145996094, 5.019000053405762, 362.94622802734375, 77.28314971923828, 336.6397705078125, 77.39521026611328], [388.3001708984375, 5.19167423248291, 452.4793701171875, 5.577062129974365, 452.38336181640625, 90.18814086914062, 387.8672790527344, 88.91836547851562], [6.535987377166748, 70.59809112548828, 156.50340270996094, 71.90779876708984, 156.2332000732422, 170.76112365722656, 6.474782943725586, 168.9591064453125], [143.13043212890625, 73.70213317871094, 225.41241455078125, 74.99711608886719, 225.60519409179688, 172.0007781982422, 143.0009002685547, 170.54078674316406], [234.8168487548828, 75.92626953125, 285.4459228515625, 77.37468719482422, 285.9234619140625, 169.67877197265625, 234.77297973632812, 168.8255157470703], [279.1575622558594, 80.0408935546875, 321.14984130859375, 81.3666000366211, 322.0708923339844, 170.91236877441406, 279.68658447265625, 170.24501037597656], [327.3981628417969, 81.63390350341797, 368.4235534667969, 83.271728515625, 368.04571533203125, 183.8792266845703, 326.7834777832031, 182.89442443847656], [391.68316650390625, 77.49728393554688, 450.87774658203125, 80.59970092773438, 450.4789123535156, 351.57135009765625, 389.3020935058594, 351.1860656738281], [18.0737361907959, 164.57754516601562, 143.79637145996094, 164.93113708496094, 144.43597412109375, 223.82362365722656, 17.937854766845703, 224.24269104003906], [143.6466522216797, 168.29611206054688, 212.39947509765625, 169.4072723388672, 213.32969665527344, 219.197021484375, 143.596923828125, 219.14364624023438], [227.44757080078125, 167.58865356445312, 278.9219970703125, 169.0375213623047, 279.76116943359375, 218.89549255371094, 227.97239685058594, 218.91354370117188], [273.3187561035156, 169.4761962890625, 319.05877685546875, 171.0519256591797, 320.08544921875, 221.45974731445312, 273.823974609375, 221.37847900390625], [316.6603088378906, 168.15016174316406, 356.101806640625, 169.28562927246094, 355.83355712890625, 223.83612060546875, 315.95855712890625, 223.8933563232422], [11.54481315612793, 209.0058135986328, 149.67161560058594, 210.10269165039062, 150.80596923828125, 313.8858642578125, 11.31093692779541, 313.5046691894531], [142.2597198486328, 220.2042999267578, 210.1415252685547, 220.74473571777344, 208.92703247070312, 306.9714660644531, 140.53517150878906, 306.3691101074219], [230.01699829101562, 218.03366088867188, 278.3423767089844, 218.091064453125, 278.7393493652344, 309.50640869140625, 229.2131805419922, 309.4779052734375], [279.83758544921875, 218.29098510742188, 322.95928955078125, 217.55064392089844, 322.50897216796875, 312.3780822753906, 278.6295471191406, 312.7047119140625], [335.04815673828125, 219.59906005859375, 366.26019287109375, 219.55929565429688, 364.4071044921875, 315.0625305175781, 332.7979736328125, 315.12701416015625]], 'html': '<html><body><table><tr><td>岗位职级名称</td><td>职级 号码</td><td>二 类城市</td><td>三类 城市</td><td>香港</td><td>国(境)外</td></tr><tr><td>总经理、副总经理、 高级专业经理、高级 销售经理、专业总监</td><td>M3及以上、 P5及以上</td><td>600</td><td>500</td><td>1500</td><td rowspan="3">标准间及 参照财行 [2013] 516号、 财行 [2017]43 4号住宿 费标准执 行</td></tr><tr><td>部门经理</td><td>M2</td><td>400</td><td>350</td><td>1100</td></tr><tr><td>专业经理、高级技师 (主任)、销售经理、 助理专业经理、技 师、助理销售经理 其余人员S1、S2、S3、P1、P2</td><td>P3、P4 S4、S5</td><td>350 300</td><td>300</td><td>1100 2601100</td></tr></table></body></html>'}, 'img_idx': 0}

已在main分支修复，将会在2.8.0版本发布

[PureWaterCatt]

@GreatV 大大main分支已经适配了吗，我新pull下来测试了一下好像没变化

[GreatV]

我是这么测试的

import os
import cv2
from paddleocr import PPStructure,draw_structure_result,save_structure_res
from PIL import Image

table_engine = PPStructure(show_log=True, image_orientation=True)

save_folder = './output'
img_path = './334352202-c065007d-96b3-4a30-ab4d-42be47ec3ee8.png'
img = cv2.imread(img_path)
result = table_engine(img)
save_structure_res(result, save_folder,os.path.basename(img_path).split('.')[0])

for line in result:
    line.pop('img')
    print(line)

font_path = 'doc/fonts/simfang.ttf' # PaddleOCR下提供字体包
image = Image.open(img_path).convert('RGB')
im_show = draw_structure_result(image, result,font_path=font_path)
im_show = Image.fromarray(im_show)
im_show.save('result.jpg')

[PureWaterCatt]

我是这么测试的

import os
import cv2
from paddleocr import PPStructure,draw_structure_result,save_structure_res
from PIL import Image

table_engine = PPStructure(show_log=True, image_orientation=True)

save_folder = './output'
img_path = './334352202-c065007d-96b3-4a30-ab4d-42be47ec3ee8.png'
img = cv2.imread(img_path)
result = table_engine(img)
save_structure_res(result, save_folder,os.path.basename(img_path).split('.')[0])

for line in result:
    line.pop('img')
    print(line)

font_path = 'doc/fonts/simfang.ttf' # PaddleOCR下提供字体包
image = Image.open(img_path).convert('RGB')
im_show = draw_structure_result(image, result,font_path=font_path)
im_show = Image.fromarray(im_show)
im_show.save('result.jpg')

好像和之前的效果是一样的，最后一行的单元格格式不对，按理来讲应该有5行

[GreatV]

是的，表格结果没有识别正确，我只关注了文字识别成功了。可能把图往外padding一些效果会好点。