Research of camera module sensor dead pixel detection

Wang, Xingheng; Gao, Shi; Liu, Yao; Cao, Jun; Deng, Xue

doi:10.1109/eeesym.2012.6258607

Cited by 7 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First, the correction light box's luminance and color temperature were assessed. Following that, light box images were taken, and gray values were compared to determine dead pixels [1]. This method can only detect large defects and has poor effect on small defects.The traditional cleaning and detection modes of image sensor stains were introduced as follows: after immersing in alcohol, the photosensitive element surface was cleaned with a dust-free wiping cloth.…”

Section: Research Statusmentioning

confidence: 99%

“…Therefore, although defect detection based on images becomes a fast and effective mean with the extensive application of machine vision technology, studies on image sensor are few. Defect detection methods for image sensor can be divided into manual detection and machine vision detection.This manual detection method [1] is not only time and labor-consuming with high labor intensity but also fails to avoid miscounts induced by visual fatigue. Since defect detection results are affected by the physical conditions and mental states of detectors, enterprises are in urgent need of an accurate and efficient automatic defect detection method.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fast defect detection of image sensor based on lightweight network

Zhang

Yue

et al. 2023

Preprint

View full text Add to dashboard Cite

Image sensor is the main device in camera production and its quality influences the imaging quality directly. Existing detection methods can not accurately detect the defects of image sensors. To address these problems, a two-branch network based on transfer learning is proposed to solve the problem of image sensor defect detection. Firstly, model parameters and detection speed were improved by replacing the backbone feature extraction network of YOLOV with a lightweight network. Secondly, anchor box optimization and focal loss were used to solve the problem of accurate positioning and unbalanced sample classification. Finally, with the cosine annealing scheduler and fixed learning rate, the network's convergence rate and overall performance are improved. The proposed lightweight network algorithm achieves a better effect than the original algorithm. The proposed algorithm has higher detection speed and smaller parameter size than other detection methods under the same size.

show abstract

Section: Research Statusmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%