Deep convolutional neural networks show a good prospect in the fertility detection and classification of specific pathogen-free hatching egg embryos in the production of avian influenza vaccine, and our previous work has mainly investigated three factors of networks to push performance: depth, width, and cardinality. However, an important problem that feeble embryos with weak blood vessels interfering with the classification of resilient fertile ones remains. Inspired by fine-grained classification, we introduce the attention mechanism into our model by proposing a dense pixelwise spatial attention module combined with the existing channel attention through depthwise separable convolutions to further enhance the network class-discriminative ability. In our fused attention module, depthwise convolutions are used for channel-specific features learning, and dilated convolutions with different sampling rates are adopted to capture spatial multiscale context and preserve rich detail, which can maintain high resolution and increase receptive fields simultaneously. The attention mask with strong semantic information generated by aggregating outputs of the spatial pyramid dilated convolution is broadcasted to low-level features via elementwise multiplications, serving as a feature selector to emphasize informative features and suppress less useful ones. A series of experiments conducted on our hatching egg dataset show that our attention network achieves a lower misjudgment rate on weak embryos and a more stable accuracy, which is up to 98.3% and 99.1% on 5-day and 9-day old eggs, respectively.
In this paper, a new convolution neural network method is proposed for the inspection and classification of galvanized stamping parts. Firstly, all workpieces are divided into normal and defective by image processing, and then the defective workpieces extracted from the region of interest (ROI) area are input to the trained fully convolutional networks (FCN). The network utilizes an end-to-end and pixel-to-pixel training convolution network that is currently the most advanced technology in semantic segmentation, predicts result of each pixel. Secondly, we mark the different pixel values of the workpiece, defect and background for the training image, and use the pixel value and the number of pixels to realize the recognition of the defects of the output picture. Finally, the defect area’s threshold depended on the needs of the project is set to achieve the specific classification of the workpiece. The experiment results show that the proposed method can successfully achieve defect detection and classification of galvanized stamping parts under ordinary camera and illumination conditions, and its accuracy can reach 99.6%. Moreover, it overcomes the problem of complex image preprocessing and difficult feature extraction and performs better adaptability.
FCN, trained end-to-end, pixels-to-pixels, predict result of each pixel. It has been widely used for semantic segmentation. In order to realize the blood vessels segmentation of hatching eggs, a method based on FCN is proposed in this paper. The training datasets are composed of patches extracted from very few images to augment data. The network combines with lower layer and deconvolution to enables precise segmentation. The proposed method frees from the problem that training deep networks need large scale samples. Experimental results on hatching eggs demonstrate that this method can yield more accurate segmentation outputs than previous researches. It provides a convenient reference for fertility detection subsequently.
On-line glue detection based on machine version is significant for rust protection and strengthening in car production. Shadow stripes caused by reflect light and unevenness of inside front cover of car reduce the accuracy of glue detection. In this paper, we propose an effective algorithm to distinguish the edges of the glue and shadow stripes. Teaching points are utilized to calculate slope between the two adjacent points. Then a tracking model based on pixel convolution along motion direction is designed to segment several local rectangular regions using distance. The distance is the height of rectangular region. The pixel convolution along the motion direction is proposed to extract edges of gules in local rectangular region. A dataset with different illumination and complexity shape stripes are used to evaluate proposed method, which include 500 thousand images captured from the camera of glue gun machine. Experimental results demonstrate that the proposed method can detect the edges of glue accurately. The shadow stripes are distinguished and removed effectively. Our method achieves the 99.9% accuracies for the image dataset.
A measuring method based on linear structured light scanning is proposed to achieve the accurate measurement of the complex internal shape of large steel plates. Firstly, by using a calibration plate with round marks, an improved line scanning calibration method is designed. The internal and external parameters of camera are determined through the calibration method. Secondly, the images of steel plates are acquired by line scan camera. Then the Canny edge detection method is used to extract approximate contours of the steel plate images, the Gauss fitting algorithm is used to extract the sub-pixel edges of the steel plate contours. Thirdly, for the problem of inaccurate restoration of contour size, by measuring the distance between adjacent points in the grid of known dimensions, the horizontal and vertical error curves of the images are obtained. Finally, these horizontal and vertical error curves can be used to correct the contours of steel plates, and then combined with the calibration parameters of internal and external, the size of these contours can be calculated. The experiments results demonstrate that the proposed method can achieve the error of 1 mm/m in 1.2m×2.6m field of view, which has satisfied the demands of industrial measurement.
During the online glue detection of body in white (BIW), the purpose of traditional glue detection based on machine vision is the localization and segmentation of glue, which is dissatisfactory for estimating the uniformity of glue with complex shape. A three-dimensional glue detection method based on the linear structured light and the movement parameters of robot is proposed. Firstly, the linear structured light and epipolar constraint algorithm are used for sign matching of binocular vision. Then, hand-eye relationship between robot and binocular camera is utilized to unified coordinate system. Finally, a structured light stripe extraction method is proposed to extract the sub-pixel coordinates of the light strip center. Experiments results demonstrate that the propose method can estimate the shape of glue accurately. For three kinds of glue with complex shape and uneven illumination, our method can detect the positions of blemishes. The absolute error of measurement is less than 1.04mm and the relative error is less than 10% respectively, which is suitable for online glue detection in BIW.
The quality of welding studs is significant for installation and localization of components of car in the process of automobile general assembly. A welding stud detection method based on line structured light is proposed. Firstly, the adaptive threshold is designed to calculate the binary images. Then, the light stripes of the image are extracted after skeleton line extraction and morphological filtering. The direction vector of the main light stripe is calculated using the length of the light stripe. Finally, the gray projections along the orientation of the main light stripe and the vertical orientation of the main light stripe are computed to obtain curves of gray projection, which are used to detect the studs. Experimental results demonstrate that the error rate of proposed method is lower than 0.1%, which is applied for automobile manufacturing.
Weakly-supervised semantic segmentation is a challenge in the field of computer vision. Most previous works utilize the labels of the whole training set and thereby need the construction of a relationship graph about image labels, thus result in expensive computation. In this study, we tackle this problem from a different perspective. We proposed a novel semantic segmentation algorithm based on background priors, which avoids the construction of a huge graph in whole training dataset. Specifically, a random forest classifier is obtained using weakly supervised training data .Then semantic texton forest (STF) feature is extracted from image superpixels. Finally, a CRF based optimization algorithm is proposed. The unary potential of CRF derived from the outputting probability of random forest classifier and the robust saliency map as background prior. Experiments on the MSRC21 dataset show that the new algorithm outperforms some previous influential weakly-supervised segmentation algorithms. Furthermore, the use of efficient decision forests classifier and parallel computing of saliency map significantly accelerates the implementation.
The majority of pedestrian detection approaches use multiscale detection and the sliding window search scheme with high computing complexity. We present a fast pedestrian detection method using the deformable part model and pyramid layer location (PLL). First, the object proposal method is used rather than the traditional sliding window to obtain pedestrian proposal regions. Then, a PLL method is proposed to select the optimal root level in the feature pyramid for each candidate window. On this basis, a single-point calculation scheme is designed to calculate the scores of candidate windows efficiently. Finally, pedestrians can be located from the images. The Institut national de recherche en informatique et en automatique dataset for human detection is used to evaluate the performance of the proposed method. The experimental results demonstrate that the proposed method can reduce the number of feature maps and windows requiring calculation in the detection process. Consequently, the computing cost is significantly reduced, with fewer false positives.
Image matching is one of the most important techniques in intelligent systems and is widely applied in many fields.
Firstly, based on integrated feature congruency, interesting target detection algorithm in complex natural backgrounds
images is studied in this paper. By detecting the abrupt changes, we can detect interesting target areas. In this paper, the
local image information is obtained by logGabor filter banks, and is represented by a collection of separate features. The
integrated features consist of some separable significant features. The integrated feature congruency model is presented
based on the integrated feature. We gain improved integrated feature congruency model by compensating noise. Then,
we get a new kind of phase-based image matching method (PIM) by combining this model and Rotation Invariant Phase
Only Correlation (RIPOC) algorithm. Experimental results show that the PIM algorithm is effective in detecting
interesting targets and locating the matching targets exactly. This algorithm is invariant to image illumination, contrast,
rotation and scaling. And this model is robust, general and accords with the human vision system (HVS) characteristics.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.