Scale Invariant and Noise Robust Interest Points With Shearlets

Improving the matching reliability of low-altitude images is one of the most challenging issues in recent years, particularly for images with large viewpoint variation. In this study, an approach for low-altitude remote sensing image matching that is robust to the geometric transformation caused by viewpoint change is proposed. First, multiresolution local regions are extracted from the images and each local region is normalized to a circular area based on a transformation. Second, interest points are detected and clustered into local regions. The feature area of each interest point is determined under the constraint of the local region which the point belongs to. Then, a descriptor is computed for each interest point by using the classical scale invariant feature transform (SIFT). Finally, a feature matching strategy is proposed on the basis of feature similarity confidence to obtain reliable matches. Experimental results show that the proposed method provides significant improvements in the number of correct matches compared with other traditional methods.

Download Full-text

ROBUST LOW-ALTITUDE IMAGE MATCHING BASED ON LOCAL REGION CONSTRAINT AND FEATURE SIMILARITY CONFIDENCE

ISPRS Annals of Photogrammetry Remote Sensing and Spatial Information Sciences ◽

10.5194/isprsannals-iii-3-19-2016 ◽

2016 ◽

Vol III-3 ◽

pp. 19-26

Author(s):

Min Chen ◽

Qing Zhu ◽

Shengzhi Huang ◽

Han Hu ◽

Jingxue Wang

Keyword(s):

Image Matching ◽

Feature Matching ◽

Local Region ◽

Geometric Transformation ◽

Interest Point ◽

Interest Points ◽

Scale Invariant ◽

Feature Similarity ◽

Low Altitude ◽

Matching Strategy

Improving the matching reliability of low-altitude images is one of the most challenging issues in recent years, particularly for images with large viewpoint variation. In this study, an approach for low-altitude remote sensing image matching that is robust to the geometric transformation caused by viewpoint change is proposed. First, multiresolution local regions are extracted from the images and each local region is normalized to a circular area based on a transformation. Second, interest points are detected and clustered into local regions. The feature area of each interest point is determined under the constraint of the local region which the point belongs to. Then, a descriptor is computed for each interest point by using the classical scale invariant feature transform (SIFT). Finally, a feature matching strategy is proposed on the basis of feature similarity confidence to obtain reliable matches. Experimental results show that the proposed method provides significant improvements in the number of correct matches compared with other traditional methods.

Download Full-text

Classifying Disease in Fruit using Machine Learning

International Journal of Scientific Research in Computer Science Engineering and Information Technology ◽

10.32628/cseit217341 ◽

2021 ◽

pp. 309-313

Author(s):

Ayeesha ◽

Fathima Zeela ◽

Vijetha

Keyword(s):

Fruits And Vegetables ◽

Detection System ◽

Early Stage ◽

Object Identification ◽

Vegetable Crops ◽

Interest Points ◽

Scale Invariant ◽

Technological Support ◽

Processing Theory ◽

Fruit Disease

India is agricultural country and Indian farmer select wide selection of fruit and vegetable crops. The cultivation of crops can be improved by the technological support. Fruits and vegetables losses are caused by disease. Diseases are seen on the leaves and fruits of plant, therefore disease detection plays a crucial role in cultivation of crops. Pathogens, fungi, microorganism, bacteria and viruses are sorts of fruit diseases also unhealthy environment is responsible for diseases. There are many techniques to spot diseases in fruits in its early stages. Hence, there's a requirement of automatic fruit unwellness detection system within the early stage of the unwellness. The aim is to detect the fruit disease, this method take input as image of fruit and determine it as infected or non- infected. The proposed method is based on the use of Scale-invariant Feature Transform (SIFT) Model with the desirable goal of accurate and fast classification of fruits. The SIFT features are local and based on the appearance of the object at particular interest points, and are invariant to image scale and rotation. They are also robust to changes in illumination, noise, and minor changes in viewpoint on image processing theory. SIFT have significant advantages because of their high accuracy, relatively easy to extract and allow for correct object identification with low probability of mismatch. Besides, they do not need an outsized number of coaching samples to avoid overfitting.

Download Full-text

Indexing based on scale invariant interest points

Proceedings Eighth IEEE International Conference on Computer Vision ICCV 2001 ICCV-01 ◽

10.1109/iccv.2001.937561 ◽

2002 ◽

Cited By ~ 466

Author(s):

K. Mikolajczyk ◽

C. Schmid

Keyword(s):

Interest Points ◽

Scale Invariant

Download Full-text

5 Detection and Measurement of Displacement and Velocity of Single Moving Object in a Stationary Background

Sir Syed Research Journal of Engineering & Technology ◽

10.33317/ssurj.v7i1.41 ◽

2018 ◽

Vol 7 (1) ◽

pp. 6

Author(s):

Amr Reda. R. Almaddah ◽

Tauseef Ahmad ◽

Abdullah Dubai

Keyword(s):

Vertical Direction ◽

Moving Object ◽

Corner Detection ◽

Interest Points ◽

Scale Invariant ◽

Current Frame ◽

Lighting Condition ◽

Harris Corner ◽

Independent Variable ◽

Harris Corner Detector

The traditional Harris detector are sensitive to noise and resolution because without the property of scale invariant. In this research, The Harris corner detector algorithm is improved, to work with multi resolution images, the technique has also been working with poor lighting condition by using histogram equalization technique. The work we have done addresses the issue of robustly detection of feature points, detected multiple of local features are characterized by the intensity changes in both horizontal and vertical direction which is called corner features. The goal of this work is to detect the corner of an object through the Harris corner detector with multiple scale of the same image. The scale invariant property applied to the Harris algorithm for improving the corner detection performance in different resolution of the same image with the same interest point. The detected points represented by two independent variables (x, y) in a matrix (x, y) and the dependent variable f are called intensity of interest points. Through these independent variable, we get the displacement and velocity of object by subtracting independent variable f(x,y) at current frame from the previous location f ̀((x,) ̀(y,) ̀) of another frame. For further work, multiple of moving object environment have been taken consideration for developing algorithms.

Download Full-text

5 Detection and Measurement of Displacement and Velocity of Single Moving Object in a Stationary Background

Sir Syed Research Journal of Engineering & Technology ◽

10.33317/ssurj.v1i1.41 ◽

2018 ◽

Vol 1 (1) ◽

pp. 6

Author(s):

Amr Reda. R. Almaddah ◽

Tauseef Ahmad ◽

Abdullah Dubai

Keyword(s):

Vertical Direction ◽

Moving Object ◽

Corner Detection ◽

Interest Points ◽

Scale Invariant ◽

Current Frame ◽

Lighting Condition ◽

Harris Corner ◽

Independent Variable ◽

Harris Corner Detector

The traditional Harris detector are sensitive to noise and resolution because without the property of scale invariant. In this research, The Harris corner detector algorithm is improved, to work with multi resolution images, the technique has also been working with poor lighting condition by using histogram equalization technique. The work we have done addresses the issue of robustly detection of feature points, detected multiple of local features are characterized by the intensity changes in both horizontal and vertical direction which is called corner features. The goal of this work is to detect the corner of an object through the Harris corner detector with multiple scale of the same image. The scale invariant property applied to the Harris algorithm for improving the corner detection performance in different resolution of the same image with the same interest point. The detected points represented by two independent variables (x, y) in a matrix (x, y) and the dependent variable f are called intensity of interest points. Through these independent variable, we get the displacement and velocity of object by subtracting independent variable f(x,y) at current frame from the previous location f ̀((x,) ̀(y,) ̀) of another frame. For further work, multiple of moving object environment have been taken consideration for developing algorithms.

Download Full-text