scholarly journals EXTRACTION METHOD FOR CENTERLINES OF RICE SEEDLINGS BASED ON FAST-SCNN SEMANTIC SEGMENTATION

2021 ◽  
pp. 335-344
Author(s):  
Yusong Chen ◽  
Changxing Geng ◽  
Yong Wang ◽  
Guofeng Zhu ◽  
Renyuan Shen
Keyword(s):  
Color Image ◽  
Semantic Segmentation ◽  
Detection Algorithm ◽  
Rice Seedling ◽  
Visual Navigation ◽  
Corner Detection ◽  
Feature Points ◽  
Rice Seedlings ◽  
Optimal Model ◽  
Outer Contour

For the extraction of paddy rice seedling centerline, this study proposed a method based on Fast-SCNN (Fast Segmentation Convolutional Neural Network) semantic segmentation network. By training the FAST-SCNN network, the optimal model was selected to separate the seedling from the picture. Feature points were extracted using the FAST (Features from Accelerated Segment Test) corner detection algorithm after the pre-processing of original images. All the outer contours of the segmentation results were extracted, and feature point classification was carried out based on the extracted outer contour. For each class of points, Hough transformation based on known points was used to fit the seedling row centerline. It has been verified by experiments that this algorithm has high robustness in each period within three weeks after transplanting. In a 1280×1024-pixel PNG format color image, the accuracy of this algorithm is 95.9% and the average time of each frame is 158ms, which meets the real-time requirement of visual navigation in paddy field.

A Self-Adaptive Corner Detection Algorithm for Low-Contrast Images

2014 ◽  
Vol 615 ◽  
pp. 158-164
Author(s):  
Liang Sun ◽  
Jian Chun Xing ◽  
Shuang Qing Wang ◽  
Shi Qiang Wang
Keyword(s):  
Large Scale ◽  
Detection Algorithm ◽  
Corner Detection ◽  
Feature Points ◽  
Low Contrast ◽  
Security Monitoring ◽  
Harris Corner ◽  
Comparative Results ◽  
Security Monitoring System ◽  
Self Adaptive

In order to effectively inhibit the image dithering caused by wind-induced vibration in the security monitoring system, it calls for the extraction and match of the feature points of the sequential frames. Harris corner detection algorithm is a widely-employed characteristics extraction algorithm in the image processing. In the security monitoring field, images and videos photographed are characterized by large scale, high pixel and low contrast degree. The classical algorithm often fails to effectively obtain the feature points while handling the images and videos of the kind. Concerning the above problems, this paper puts forward an improved self-adaptive corner detection algorithm. Firstly, this paper employs the self-adaptive gray threshold comparative results of the of every point with the surrounding eight neighborhood points to select the preselected points of part of the corners. Following that, this paper classifies the preselected points into three types according to certain rules and the value of the already selected self-adaptive gray threshold. At last, according to the classification results, this paper uses different corners to test function threshold and the preselected points as well to eliminate the peripheral points and the pseudo-corners so as to gain the genuine corners. After verifying the above improved algorithm in the practical scenario in the security monitoring, the results of this paper prove its effectiveness, feasibility and its advantages in terms of robustness.

Yawning Detection Based on Mouth Feature Points Curve Fitting

2012 ◽  
Vol 605-607 ◽  
pp. 2227-2231
Author(s):  
Wu Yang Ding ◽  
Ling Zhang ◽  
Yun Hua Chen
Keyword(s):  
Mathematical Model ◽  
Real Time ◽  
Curve Fitting ◽  
Detection Method ◽  
Detection Algorithm ◽  
Corner Detection ◽  
Feature Points ◽  
Harris Corner ◽  
Fatigue Monitoring ◽  
Harris Corner Detection

A yawning detection method which can be used in drivers’ fatigue monitoring is proposed. To adapt to the variance in different mouth shapes and sizes, it based on mouth inner contour corner detection and curve fitting. First, the Harris corner detection algorithm was used to detect inner mouth feature points. Second, we established the open mouths’ mathematical model by curve fitting those points, calculated the degree of mouth openness using the mouth model, and generated the real-time M-curve. Third, the duration of big openness in successive images is divided into levels for further judgment. The validation results show that the method can obtain more precise mouth parameters and distinguish yawn from complex mouth activities. So the method achieves a higher level of accuracy.

Image Template Matching Algorithm for Removing Useless Feature Points

2011 ◽  
Vol 271-273 ◽  
pp. 19-23 ◽  
Author(s):  
Jian Shu Gao ◽  
Tao Yang ◽  
Zhi Jing Yu
Keyword(s):  
Template Matching ◽  
Detection Algorithm ◽  
Corner Detection ◽  
Image Feature ◽  
Feature Points ◽  
Illumination Variation ◽  
Matching Algorithm ◽  
Plane Image ◽  
Special Points ◽  
Fixed Structure

With the aim to remove the unwanted feature points, template matching algorithm which includes pixel-coherence and fixed structure is proposed in this paper. Corner detection algorithm can extract image feature points with the flexibility for illumination variation and affine transformation. This paper gets the special points of plane image by corner detection algorithm and proposes the problem of existing more useless point. Then, template matching algorithm is used to solve the issue and collects the best match points. Finally, the paper compares the template image and the matched image. The experimental results prove that the feature points by template matching algorithm have excellent accurate characteristic for the illumination, transfer and the rotation transform, and they are highly satisfactory.

Moving Object Detection and Tracking Using Particle Filter

2013 ◽  
Vol 321-324 ◽  
pp. 1200-1204 ◽  
Author(s):  
M.M. Naushad Ali ◽  
M. Abdullah-Al-Wadud ◽  
Seok Lyong Lee
Keyword(s):  
Particle Filter ◽  
Detection Algorithm ◽  
Human Motion ◽  
Human Detection ◽  
Corner Detection ◽  
Feature Points ◽  
Detection And Tracking ◽  
Harris Corner ◽  
Challenging Tasks ◽  
Non Gaussian

Moving human detection and tracking are challenging tasks in computer vision. Human motion is usually non-linear and non-Gaussian, and thus many common algorithms are not appropriate for tracking. In this paper we propose a robust tracking algorithm based on particle filter. Multiple moving human in a video sequence are detected using frame difference and morphological operation. Then feature points of every person are extracted using a Harris Corner detection algorithm. Finally, Histogram of Oriented Gradient (HOG) is calculated for each feature point and feature points of the corresponding person are tracked using particle filter. Experimental results demonstrate that our method is efficient to improve the performance of tracking.

scholarly journals Segmentation of Rice Seedlings Using the YCrCb Color Space and an Improved Otsu Method

Agronomy ◽  
2018 ◽  
Vol 8 (11) ◽  
pp. 269 ◽  
Author(s):  
Juan Liao ◽  
Yao Wang ◽  
Junnan Yin ◽  
Lu Liu ◽  
Shun Zhang ◽  
...  
Keyword(s):  
Color Space ◽  
Computational Cost ◽  
Rice Seedling ◽  
Visual Navigation ◽  
Paddy Fields ◽  
Rice Seedlings ◽  
Illumination Variation ◽  
Search Range ◽  
Otsu Method ◽  
Green Index

Rice seedling segmentation is a fundamental process of extracting the guidance line for automated rice transplanters with a visual navigation system, which can provide crop row information to ensure the transplanter plants seedlings along the crop row without damaging seedlings. However, obtaining accurate rice seedling segmentation in paddy fields is still a challenging task. In this paper, a rice seedling segmentation method in paddy fields is proposed. The method mainly consists of two steps: image graying and threshold segmentation. In the procedure of image graying, the RGB (Red Green Blue) seedling image is first converted into the YCrCb color space and a Cg component is constructed. A color-index 2Cg-Cb-Cr is then constructed for image graying based on the excess green index (2G-R-B), which can reduce the influence of illumination variation on the equality of image graying. For the second step, an improved Otsu method is proposed to segment rice seedlings. With respect to the improved Otsu method in this research, the background variance of within class variance is weighted by a probability parameter to ensure that the method works well for both bimodal and near-unimodal histogram images, and the search range of gray levels is constrained to reduce the time to search the segmentation threshold. Experimental results indicate that the proposed method achieves better segmentation results and reduces the computational cost compared with the traditional Otsu method and other improved Otsu methods.

Feature Extraction Method Based on Moment Invariants

2014 ◽  
Vol 936 ◽  
pp. 2263-2266
Author(s):  
Wan Bing Li ◽  
Hong Wei Quan ◽  
Xia Fei Huang
Keyword(s):  
Detection Algorithm ◽  
Corner Detection ◽  
Reference Image ◽  
Feature Points ◽  
Moment Invariant ◽  
Feature Extraction Method ◽  
Harris Corner ◽  
Initial Control ◽  
Image Set ◽  
Harris Corner Detection

To match two or more images originated from the same scenario, a new fast automatic registration algorithm based on sparse feature point extraction is proposed. At the first step, the improved Harris corner detection algorithm is used to get two sets of feature points from the reference image and registration image. Second, a group of sparse feature points are selected from the reference image set as initial control points. Then, the corresponding matching points in the registration image set are searched based on local moment invariant similarity detection. Experimental results demonstrate that this method is fast and efficient.

scholarly journals Real-time 2D–3D door detection and state classification on a low-power device

2021 ◽  
Vol 3 (5) ◽  
Author(s):  
João Gaspar Ramôa ◽  
Vasco Lopes ◽  
Luís A. Alexandre ◽  
S. Mogo
Keyword(s):  
Low Power ◽  
Real Time ◽  
Object Classification ◽  
Semantic Segmentation ◽  
Detection Algorithm ◽  
Power Device ◽  
Indoor Environments ◽  
State Classification ◽  
Segmentation Algorithms ◽  
Indoor Spaces

AbstractIn this paper, we propose three methods for door state classification with the goal to improve robot navigation in indoor spaces. These methods were also developed to be used in other areas and applications since they are not limited to door detection as other related works are. Our methods work offline, in low-powered computers as the Jetson Nano, in real-time with the ability to differentiate between open, closed and semi-open doors. We use the 3D object classification, PointNet, real-time semantic segmentation algorithms such as, FastFCN, FC-HarDNet, SegNet and BiSeNet, the object detection algorithm, DetectNet and 2D object classification networks, AlexNet and GoogleNet. We built a 3D and RGB door dataset with images from several indoor environments using a 3D Realsense camera D435. This dataset is freely available online. All methods are analysed taking into account their accuracy and the speed of the algorithm in a low powered computer. We conclude that it is possible to have a door classification algorithm running in real-time on a low-power device.

An Improved Corner Detection Algorithm Based on Harris

2018 ◽  
Author(s):  
Songqi Han ◽  
Weibo Yu ◽  
Hongtao Yang ◽  
Shicheng Wan
Keyword(s):  
Detection Algorithm ◽  
Corner Detection
Sign in / Sign up

Export Citation Format

Share Document