Batch image recognition form: convert multiple images at once

Key points of this article

OmniDoc's OCR is processed by file in the PDF editor of the desktop client. Multiple images can be merged into one PDF first and then recognized.
Arrange the pictures in table order before merging, and make the text on each page clear, correct, and not blocked, which can reduce misaligned rows and columns later.
After identifying and exporting Excel, you need to check the numbers, merged cells and table headers page by page, especially the decimal point, negative sign and date.
The identification result is only a draft. When it involves reimbursement, financial or contract figures, it should be compared with the original picture cell by cell before use.

I have more than a dozen photos of meeting minutes, monthly reports or purchase lists in my hand. It is necessary to turn them all into Excel that can be filtered and summed. Manually entering them one by one is slow and error-prone. Many people search for "Batch recognition of text from images to tables" and "Batch recognition of photos from tables". In fact, what they want to solve is the same thing: how to process multiple pictures at one time instead of repeating the complete operation for each picture.

This article is written for people who need to organize multiple paper forms, scans or screenshots, such as administrators, financial colleagues, students and scientific researchers. After reading, you will know that OmniDoc's OCR processes files on a file-by-file basis, so the correct approach is to first merge multiple pictures into a PDF, then recognize and export them all at once, rather than uploading them one by one. The article will also explain how to check after identification so that you can use the results with confidence.

How to convert multiple table images into Excel at one time? One sentence conclusion

First merge the table images of the same batch into a PDF in order, then perform OCR in the PDF editor of the OmniDoc desktop client, and finally export to Excel. This only requires one recognition and one export, which is more trouble-free than converting one picture at a time, and also facilitates unified verification.

your situation	Recommended practices	reason
Mobile phone photo of a batch of paper forms	First merge into a PDF, then do OCR and export to Excel	Most of the photos are scanned images, and OCR needs to be processed in PDF, which can be completed in one go after merging.
Computer screenshots or exported table images	Also merge into PDF, focus on checking column alignment after recognition	The resolution of screenshots is usually higher, but the column widths of multiple screenshots may be inconsistent.
There are multiple small tables in one picture	Split the table into separate pages or separate pictures and then merge them	Different tables are mixed on one page and can be easily serialized or concatenated after identification.
Larger numbers, used for reimbursement or finance	After recognition, compare it with the original image one by one, and then use it for calculation.	Numbers are error-prone and must be confirmed manually
Tables are Word or PDF files, not images	Export to Excel directly with PDF editor without OCR	The text itself can be selected, no recognition is required, and the result is more stable.

Before you start: Confirm these points first

Spend a few minutes preparing before you start, and you'll save a lot of rework later on.

Confirm whether the picture really needs to be recognized. If the text in the PDF can be selected with the mouse, it means it is already a text layer and can be exported directly. The judgment method is simple: open the file, try to select a piece of text and copy it. If not selected, it is a scan or picture that requires OCR.
Organize pictures in table order. The page order after merging is the top-down order in Excel. Numbering the files with file names in advance, such as "01, 02, 03", can avoid being reversed after exporting.
Check the clarity of each picture. For parts that are blurred, tilted, reflective or shaded, the recognition error rate will increase significantly. When shooting, try to face the table directly, with even lighting, so that the entire table is within the frame.
Unify the orientation of pictures. If the horizontal table is photographed vertically, the rows and columns will be misaligned during recognition. Rotate each image to the correct orientation before merging.
Original pictures ready. The recognition results need to be compared with the original images, so do not delete the originals and keep a complete image folder.

Steps to convert multiple table images to Excel in OmniDoc

OmniDoc's OCR function is located in the PDF editor of the desktop client and processes a single file, so multiple images need to be merged first.

Merge images into one PDF. You can open the image on your computer, use the system's print function, change the printer to a virtual PDF output option and save it, or use common image to PDF tools on mobile phones and computers. Each figure occupies one page, and the page order is consistent with the table order.
Download and open the OmniDoc desktop client. Go to download.omnidoc.top to download the version corresponding to your system (Windows, macOS or Linux), install and open it.
Open the merged file with a PDF editor. After opening the file, read it page by page to make sure that each page is a complete table and there are no missing or duplicate pages.
Perform OCR on the scanned content. Select the OCR function in the PDF editor to let the software recognize the text on the page. The entire document will be processed together, there is no need to work on each page individually.
Wait for the recognition to be completed and then browse page by page. Check whether the recognized text is consistent with the original image, paying special attention to the numbers, decimal points and headers. If you find problems, you can modify the paragraph text directly in the PDF editor.
Export to Excel. Select the option to export to Excel to save locally. Write down the file names before and after exporting to facilitate distinguishing versions later.
Open the results in Excel. Check whether the rows and columns are consistent with the original table, adjust the column width if necessary, and add filters to the table header to facilitate subsequent use.
Save the final version. Save the checked Excel copy in the same folder as the original image for easy tracing in the future.

Check each item after completion

The result of OCR is a draft, not an equivalent copy of the original. It is recommended to go through the following check items one by one.

Check items	How to see it	What should you do if there's a problem?
Whether the numbers match	Randomly select a dozen or so numbers and compare them with the original image, focusing on the decimal point, minus sign, and thousandth comma	Fix it directly in Excel and go back to check if similar numbers have the same error
Whether the rows and columns are aligned	Check whether each row of cells falls under the correct column and if there are whole row sequences	If a page is misaligned as a whole, usually the image is tilted; retake or rotate the page before identifying it
Table header and merge cells	Check whether titles spanning rows and columns have been split or missing	Manually merge or fill in table headers, and if necessary, organize them into a unified format in Excel
Date and serial number	Date format, order number, ID number, etc., are easily recognized as other characters	Check each item carefully; errors in these fields have significant consequences and should be fully confirmed with the original drawing
Check if the page is complete	Compare the number of pages with the table to ensure no missing pages, duplicate pages, or blank pages	Return to the merged PDF supplement or delete page, then re-identify
Chinese characters and proper nouns	Check whether personal names, place names, and organization names contain errors in similar characters	By modifying according to the original image, commonly used names can be uniformly handled in Excel using lookup and replacement

When verifying, remember that OCR is machine recognition, and results can be mistaken, especially in areas that are blurry, handwritten, or obscured by stamps. Data involving finance, contracts, or formal submissions should be compared item by item with the originals and not directly adopted.

📊 Core Comparison

Images from different sources have different main risks and focus areas. Below is a comparison based on the image sources, helping you decide which step to focus your energy on.

Image source	Major risks	Key Points of Processing
Paper forms taken by phone photos	Perspective distortion, shadows, and reflections can easily cause rows and columns to be misaligned	Before recognition, tap the correct and correct direction; After recognition, focus on checking column alignment
Screenshots taken from computers or exported table images	Multiple screenshots have inconsistent column widths and font sizes, resulting in uneven formats after export	Clarity is usually good, with the key to standardizing column width and header format after export
PDF that already allows selection of Chinese text	There are almost no recognition errors, but there may be residual copied formatting	No need for OCR—just export directly to Excel and check the format

❓ Frequently Asked Questions (FAQ)

1. Can OmniDoc upload many images at once and convert them to Excel?

OmniDoc's OCR function processes images by file. It is recommended to first merge multiple images into one PDF, then uniformly recognize and export them. This process can be completed in one go.

2. Can table photos taken with a phone be used directly?

You can first merge it into a PDF before recognizing it, but the clarity, angle, and lighting of the photo will directly affect the result. When shooting, try to align the table with even lighting; after identification, check each item individually.

3. What should I do if the table becomes a column of text after recognition?

This usually indicates that the structure of the rows and columns has not been properly restored. You can first check whether the original image is skewed or has a low resolution, then retake the photo if necessary, re-identify the page, or manually adjust the columns in Excel.

4. Is there a way to directly convert multiple images into a table without merging?

If you only need text from the image, you can recognize each PDF or image on the desktop and then manually summarize it. However, bulk organizing multiple images into one Excel file and merging them into PDF is usually more reliable and easier to verify for unified processing.

5. Can forms made using this method be used directly for reimbursement or submission?

It is not recommended to use them directly. OCR results need to be checked item-by-item with the original, especially the amount, date, and number. Only after confirming everything is correct before reimbursement or submission can you avoid trouble caused by identification errors.

📝 Summary

The key to converting multiple table images to Excel is to first merge them and then process them uniformly: create a PDF in order, use OCR recognition in the OmniDoc desktop client, and export it to Excel. After recognition, be sure to check each number, row, column, header, and date to ensure the results match the original image.

Open OmniDoc to identify and translate the scan