Papers by Keyword: Image Processing

Abstract: This research presents the result of study and development of image processing program to monitor the copper wire placement process prior to the autonomous soldering process. The program has been developed using Visual Studio C++6.0 with Open CV, which is software that is on the command image processing. The experiment has 3 parameters to consider, focal length of camera, color of lighting and lighting control the pieces. The experiment shows the result that control ambient light improve the image processing output. Using focal length of 100 mm gives error of less using focal length of 130 mm. The light illuminated color affects the reflection, red-light gives low reflect, compare to white-light. The image processing using this setup can detect soldering point up to 98.5 % and 92.5 % of copper wire is detected. The setup using white-light cannot detect soldering coordinates, and able to detect copper wire only 42.5%. To control the ambient light, the focal length of 100 mm, and the red illuminating suited for the image processing system of the autonomous soldering and copper wire placement.

Abstract: Since seeds are the foundation of agriculture and the Thai government plans to make Thailand an export hub of seeds under the Asean Economic Community (AEC) 2015, seed quality plays an important role in the seed production. Traditionally, physical attributes of seeds are inspected by human. However this method is very time-consuming and it highly relies on human skills and experience. Thus, in this paper, we focus on seed quality inspection of sweet pepper seeds using image processing techniques. Sweet peppers are very interesting since they have been one of the most important vegetable around the world and they have a variety of vitamins and nutrients. To identify defective sweet pepper seeds, two features used in our proposed algorithm are seed color and seed size. As shown in the results, percent accuracy of abnormal seed color and unaccepted seed size detection are 95.82% and 90.76%, respectively.

Abstract: The classification of ground-based cloud images has received more attention recently. The result of this work applies to the analysis of climate change; a correct classification is, therefore, important. In this paper, we used 18 texture features to distinguish 7 sky conditions. The important parameters of two classifiers are fine-tuned in the experiment, namely, k-nearest neighbor (k-NN) and artificial neural network (ANN). The performances of the two classifications were compared. Advantages and limitations of both classifiers were discussed. Our result revealed that the k-NN model performed at 72.99% accuracy while the ANN model has higher performance at 86.93% accuracy. We showed that our result is better than previous studies. Finally, seven most effective texture features are recommended to be used in the field of cloud type classification.

Abstract: To investigate the unsaturated soil properties, the measurement of volume change is dealt with both water and air volume changes. It cannot be defined by using only the volume of water flown in out from the specimen as saturated soil. This paper presents the volume change measurement by using image processing (IP) from the series of consolidated undrained (CU) and constant water content (CW) tests for saturated and unsaturated conditions of soil, respectively. The results, obtained between IP and the cell liquid measurement considered movement of piston and the cell expansion effect with pressure including creep, are good corresponding except the specimen has an irregular deformation.

Abstract: This article studies image copyright protection algorithm based on watermark. Aimed at the defect that watermark loading algorithm easily leads to the loss of original image information, it proposes a kind of quantization based on wavelet tree structure so that in the watermark extraction process, it can restore the original intact image without original image or watermark information. First, it makes use of level-four wavelet transform to construct tree wavelet coefficient waveband diagram. Then it discretizes and insets watermark information into different waveband nodes of tree structure according to coefficient variation maximum principle. By designing self-adaption threshold value during extracting process, it can restrict some common watermarking loss better, such as smoothing, mean value filtering and gaussian noise. The experiment result proves that this method can make watermark image the visually same as the original image and watermark can still be extracted effectively after preprocessing.

Abstract: Proposed on the basis of nondestructive testing technology, such as ultrasonic testing and X-ray detecting, image processing of hull welding based on MATLAB is a simple detecting method to detect the quality of welding. By image processing, physical dimensions, such as circumferences and shapes, as well as the intensity and texture features of welds can be concluded. Then, through numerical analysis, quality problems, including porosities, slags, undercuts, overlaps, welding defects and others, if exist, can be determined. This method decreases the cost and reduces the harm to people’s health. Experimental results showed that the detecting system can calculate and determine the major features of the welds effectively and accurately, and possesses good practical value.

Abstract: For the number of buildings in the most commonly forged composite image blur operation, this paper presents an image forensics morphological filtering algorithm based on edge characteristics; The method uses homomorphic filtering and morphological filtering method to enhance the operation of the fuzzy image edge analyzed defocus blur and artificial blurred boundary characteristics, the use of artificial edge blur and defocus blur characteristics achieved for passive blind digital image forgery evidence; experiments show that this method can effectively detect forged digital images fuzzy operation marks and the partial image has been able to accurately locate the fuzzy operation.

Abstract: This paper focuses on the visual tracking algorithm in optical imaging surveillance and tracking system. The tracking particle filter framework deemed find sparse representation problem, can effectively overcome the visual tracking algorithm appears in noise, occlusion, background interference and complex situations such as illumination changes. Morphological methods using digital occlusion area is detected to determine whether the date is added to the template tracking results set, thereby updating the control template, to effectively prevent the drift tracking results.

Abstract: Based on the shape of the image retrieval occupy an important position in the content-based image retrieval, and studied architecture, content-based image retrieval system, ie research-based image retrieval key technologies shape features for image noise in addition to the morphological processing; image segmentation; shape-based feature extraction and regional boundaries and description techniques and similarity measure techniques. The results show that the algorithm can effectively identify the characteristics of the image.

Abstract: Optical information processing systems have been widely applied, this paper studies the calculation of the amount of a reduction in the use of genetic algorithms to increase the diversity of the sample algorithm to reduce problems caused by the sample dilution with a resampling method, combined with the finite field resampling, resampling constitute a finite field of genetic algorithms. We using surface to capture the dynamics of template update, to prevent inappropriate tracking results into the template set, pending the outcome of the tracking template is added to the first test for cover. Detect occlusion control template updates. Prevent tracking results of drift.