Vehicle Quality Inspection Report Process System
The client is a leading U.S. commercial bus manufacturer in North America and produces about 23 buses a week. The external framework of each bus is assembled in Canada, and the internal parts are assembled in the United States. In order to ensure the quality, assembly line workers check vehicles, and handwrite a description of all differences and error codes on a checklist. Therefore, each bus has a 42-page quality inspection report. In order to manage and analyze the statistical information of inspection data, the handwritten inspection data had to be entered into a uniform Excel spreadsheet in a timely manner.
Therefore, the customer wants to use OCR / ICR software to recognize handwritten data and save into Excel tables to save time and cost.
Based on the analysis of the customer’s image samples, it was determined although OCR / ICR technology can recognize some of the machine printed and hand-written characters, the following problems existed:
- Handwritten styles vary from individual to individual. It was often unpredictable and even difficult for the human eye to identify
- Production environment affected sample quality. There is a lot of noises such as ink, smudges, illegible marks, which seriously affects recognition results.
After preliminary requirements analysis, aimed to address the above problems, we first helped our client understand that pure technology couldn’t solve the problem. The best solution was BPO + IT technology, which is human services and software completing the project together. The client agreed that a BPO + IT solution was the best approach and agreed to give it a try.
BPO + IT Solutions:
- Customize software entry tool based on OCR, ICR and IT technology
- BPO team complete data entry and proof.
Main workflow is as below:
- Client collected vehicle inspection reports, scanned and uploaded pictures to FTP server.
- We download images, used customized entry software to complete data input, proofed and uploaded the results to the FTP server.
- Client downloaded, approved results and provided feedback.
Work Order Recognition
In long-distance transportation, drivers often become fatigued increasing accident rates. To reduce this loss, the US DOT regulates freight drivers’ continuous driving time to ensure the drivers do not exceed drive time limits. For management, long-distance drivers are required to strictly fill out driving logs every day, using lines to record their daily schedule on a graduated scale, for example, time spent driving, eating, sleeping, etc. It was labor consuming to enter this information into a database the loss of a unified database. In order to improve efficiency, the customer wanted to use OCR to process the data.
OCR Application and Problems
OCR is generally for character recognition, but the entry information is shown via a series of lines, and is essentially simpler than character recognition. We did not utilize our own character recognition RTK, but developed new software to meet the customer’s needs.
We developed an Online Driver Log Recognition System to allow client to recognize this driver logs online.