Photometric stereo-based defect detection for lithium battery top covers

Detecting surface defects on lithium battery top covers is critical for ensuring their performance, reliability, and longevity. Photometric stereo (PS) methods, known for their ability to capture detailed surface topography with high precision and speed, have been widely applied in industrial defect detection. However, these methods often face challenges in dynamic scenes with varying lighting conditions and in detecting subtle defects such as shallow scratches. To address these limitations, we propose an enhanced approach that combines the advantages of photometric stereo with a line-scan camera and weighted least squares filtering.

In the broader publication record, this work sits in Seventh International Conference on Image Processing and Machine Vision (IPMV 2025), vol. 13636, 100-107 and connects to practical problems in 3D sensing, computational geometry, and industrial machine vision.

Problem setting

Detecting surface defects on lithium battery top covers is critical for ensuring their performance, reliability, and longevity. Photometric stereo (PS) methods, known for their ability to capture detailed surface topography with high precision and speed, have been widely applied in industrial defect detection. However, these methods often face challenges in dynamic scenes with varying lighting conditions and in detecting subtle defects such as shallow scratches.

In the broader publication record, this work appears in Seventh International Conference on Image Processing and Machine Vision (IPMV 2025), vol. 13636, 100-107. The visual notes below pair the paper’s original figures with a concise reading of the method, experimental setup, and reported results.

Method and visual evidence

The method combines domain-specific measurements with an algorithmic representation that exposes the relevant structure, then refines it into a reconstruction, correspondence, segmentation, measurement, or decision result.

The extracted figures below show the main pipeline and representative experimental evidence.

Method overview. This image is extracted from an embedded PDF image object on page 3, then recomposed for web display.

Representation and setup. This image is extracted from an embedded PDF image object on page 3, then recomposed for web display.

Experimental evidence. This image is extracted from an embedded PDF image object on page 5, then recomposed for web display.

Result comparison. This image is extracted from an embedded PDF image object on page 6, then recomposed for web display.

Additional visual result. This image is extracted from an embedded PDF image object on page 7, then recomposed for web display.

Results and impact

The evaluation reported in Seventh International Conference on Image Processing and Machine Vision (IPMV 2025), vol. 13636, 100-107 uses the extracted figures above to show the method’s measurement, reconstruction, segmentation, matching, or diagnostic behavior on representative experiments. These visuals are paired with the paper’s quantitative or qualitative analysis to make the workflow easier to inspect from the homepage.

Source handling

I extracted 9 candidate image objects from paper.pdf and generated the compressed WebP figures used on this page.