Noise Blobs Research Articles

An efficient zero-labeling segmentation approach for pest monitoring on smartphone-based images

Timely and precise farm inspection, which involves the identification and recognition of harmful insects and diseases, is crucial for safeguarding crop production. Traditional vision-based pest recognition methods typically require extensive annotated data for each pest species and a lengthy training process. This approach is time-consuming, labor-intensive, and prone to human error. Zero-shot learning offers a potential solution by enabling pest segmentation and control without requiring explicit training data. This study supports farmers in automatically identifying ten common pests and their precise locations in real-world outdoor environments. The zero-shot pest segmentation is based on a hybrid approach combining Explainable Contrastive Language-Image Pre-training (ECLIP) and Segment-Anything (SAM). Moreover, an optimized super-resolution model and various data augmentation methods are implemented to improve the quality of the dataset. Lastly, a mask post-processing step is applied to remove highly overlapping segmented masks and noise blobs caused by the complex background. The mean Intersection over Union (mIoU) of 66.5 % on the validation set demonstrates the potential of zero-shot methods for automated pest segmentation during farm inspections. This research lays the foundation for accurate pest monitoring systems capable of adapting to new pests, ultimately improving agricultural productivity.

Vision-Based In-Flight Collision Avoidance Control Based on Background Subtraction Using Embedded System.

The development of high-performance, low-cost unmanned aerial vehicles paired with rapid progress in vision-based perception systems herald a new era of autonomous flight systems with mission-ready capabilities. One of the key features of an autonomous UAV is a robust mid-air collision avoidance strategy. This paper proposes a vision-based in-flight collision avoidance system based on background subtraction using an embedded computing system for unmanned aerial vehicles (UAVs). The pipeline of proposed in-flight collision avoidance system is as follows: (i) subtract dynamic background subtraction to remove it and to detect moving objects, (ii) denoise using morphology and binarization methods, (iii) cluster the moving objects and remove noise blobs, using Euclidean clustering, (iv) distinguish independent objects and track the movement using the Kalman filter, and (v) avoid collision, using the proposed decision-making techniques. This work focuses on the design and the demonstration of a vision-based fast-moving object detection and tracking system with decision-making capabilities to perform evasive maneuvers to replace a high-vision system such as event camera. The novelty of our method lies in the motion-compensating moving object detection framework, which accomplishes the task with background subtraction via a two-dimensional transformation approximation. Clustering and tracking algorithms process detection data to track independent objects, and stereo-camera-based distance estimation is conducted to estimate the three-dimensional trajectory, which is then used during decision-making procedures. The examination of the system is conducted with a test quadrotor UAV, and appropriate algorithm parameters for various requirements are deduced.

Unsupervised Recognition and Characterization of the Reflected Laser Lines for Robotic Gas Metal Arc Welding

Unsupervised recognition of the reflected laser lines from the arc-light-modified background is prerequisite for the subsequent measurement and characterization of the weld pool shape, which is of great importance for the modeling and control of robotic arc welding. To facilitate the unsupervised recognition, the reflected laser lines need to be segmented as accurate as possible, which requires the segmented laser lines to be as continuous as possible to decrease the adverse effect of the noise blobs. In this paper, the intensity distribution caused by the arc light in the captured image is modeled. Based on the model, an efficient and robust approach is proposed, and it comprises six parts: reduction of the uneven image background by a difference operation, spline enhancement to remove the fuzziness, a gradient detection filter to eliminate the uneven background further, segmentation by an effective threshold selection method, removal of the noise blobs adaptively, and clustering based on the online computed slope of the laser line. After the laser line is clustered, a second-order polynomial is fitted to it. Finally, the weld pool is characterized by the parameters of the clustered laser line and its fitted polynomial. Experimental results verified that the proposed approach for unsupervised reflected laser line recognition is significantly superior to the state-of-the-art approach in terms of recognition accuracy.

Generalizing cell segmentation and quantification

BackgroundIn recent years, the microscopy technology for imaging cells has developed greatly and rapidly. The accompanying requirements for automatic segmentation and quantification of the imaged cells are becoming more and more. After studied widely in both scientific research and industrial applications for many decades, cell segmentation has achieved great progress, especially in segmenting some specific types of cells, e.g. muscle cells. However, it lacks a framework to address the cell segmentation problems generally. On the contrary, different segmentation methods were proposed to address the different types of cells, which makes the research work divergent. In addition, most of the popular segmentation and quantification tools usually require a great part of manual work.ResultsTo make the cell segmentation work more convergent, we propose a framework that is able to segment different kinds of cells automatically and robustly in this paper. This framework evolves the previously proposed method in segmenting the muscle cells and generalizes it to be suitable for segmenting and quantifying a variety of cell images by adding more union cases. Compared to the previous methods, the segmentation and quantification accuracy of the proposed framework is also improved by three novel procedures: (1) a simplified calibration method is proposed and added for the threshold selection process; (2) a noise blob filter is proposed to get rid of the noise blobs. (3) a boundary smoothing filter is proposed to reduce the false seeds produced by the iterative erosion. As it turned out, the quantification accuracy of the proposed framework increases from 93.4 to 96.8% compared to the previous method. In addition, the accuracy of the proposed framework is also better in quantifying the muscle cells than two available state-of-the-art methods.ConclusionsThe proposed framework is able to automatically segment and quantify more types of cells than state-of-the-art methods.

Automatic segmentation and classification of the reflected laser dots during analytic measurement of mirror surfaces

In the past research, we have proposed a one-shot-projection method for analytic measurement of the shapes of the mirror surfaces, which utilizes the information of two captured laser dots patterns to reconstruct the mirror surfaces. Yet, the automatic image processing algorithms to extract the laser dots patterns have not been addressed. In this paper, a series of automatic image processing algorithms are proposed to segment and classify the projected laser dots robustly and efficiently during measuring the shapes of mirror surfaces and each algorithm is indispensible for the finally achieved accuracy. Firstly, the captured image is modeled and filtered by the designed frequency domain filter. Then, it is segmented by a robust threshold selection method. A novel iterative erosion method is proposed to separate connected dots. Novel methods to remove noise blob and retrieve missing dots are also proposed. An effective registration method is used to help to select the used SNF laser and the dot generation pattern by analyzing if the dot pattern obeys the principle of central projection well. Experimental results verified the effectiveness of all the proposed algorithms.

Automatic navigation path detection method based on machine vision for tillage machines working on high crop stubble fields.

Abstract: Due to the influence of complex working environment and artificial factors, it is easy to cause crop up over or less tillage problem when straw returning machine is working in paddy field. A new method for path detection suitable for rice, rape and wheat high crop stubble tilling environments was proposed. First the distribution characteristics of rice, rape and wheat high crop stubble images in paddy field based on RGB color model were analyzed, and rice, the color images of rape and wheat high crop stubble were converted into gray ones using custom factor combination R+G-2B; Then, the gray images of rice, rape and wheat high crop stubble were segmented from soil background by means of luminance mean texture descriptor; Next, the binary image through custom shear-binary-image algorithm was cut to remove big noise blobs in high crop stubble’s tilled area; Finally, navigation path from navigation points by using the least square method was derived. The experimental results indicated that the navigation path detection algorithm was fast and effective to obtain navigation path in rice, rape and wheat high crop stubble tilling environments with up to 96.7% of segmentation accuracy within 0.6 s of processing time. Keywords: high crop stubble, paddy field tilling, texture statistics, road navigation, vision navigation DOI: 10.3965/j.ijabe.20140704.004 Citation: Zhang T, Xia J F, Wu G, Zhai J B. Automatic navigation path detection method of high crop stubble in paddy field tilling based on machine vision. Int J Agric & Biol Eng, 2014; 7(4): 29－37.

A Method for Detection and Removal of Rain in Videos

摘要: 降雨天气往往导致监控视频质量下降. 本文提出首先在对数图像处理(Logarithmic image processing, LIP)框架下利用灰色调约束检测出候选雨滴,进而利用主成分分析(Principal component analysis, PCA)方法计算每个候选雨滴的倾斜方向并构建其概率密度分布函数,利用Mean-shift算法估计该分布函数的峰值,作为检测到的雨滴降落方向,然后,通过方向约束去除候选雨滴中的干扰噪声. 最后,文章采用一种加权的重构方法消除雨滴. 实验证明,算法能够有效检测并去除各种场景中的雨滴. 关键词: 对数图像处理 / 均值漂移 / 主成分分析 / 雨滴消除

A Robust Location Algorithm for PCB's Solder Joints

Printed circuit boards (PCBs) are being widely used in the electronic packages. Solder joints are often used to interconnect chip resistors and other components onto PCBs. The defects of solder joints will increase quality costs and deteriorate performance. If solder joints cant be located accuracy, AOI system will not inspect solder joints at the right place of solder joints and must lead to misjudgement. Especially, the misjudgement will increase dramatically under uncertain noise disturbance to micro-size solder joints due to inaccuracy location. In this work, an eliminating uncertain noise method is proposed and a robust location algorithm for PCBs solder joints is present. Firstly, some location windows are set based on technology parameters of chip and its solder joints, and solder joints feature image is obtained from it solder joint image based on series of image pre-processing. Secondly, the layout frame outside solder joints, which is viewed as a noise disturbance to location, is extracted as a binary image and is projected to X axis and Y axis, then the smaller region, which include solder joints but no the layout frame, is obtained based on calculating the sum features function of layout frame. Thirdly, the blob feature image of solder joints is extracted from its gray images of red, green, blue layers; it may have some noise blob around solder joints feature, then an evaluation function is develop to judge blob, which maybe belongs to noise or solder joints feature, and only solder joints feature are remained. Fourthly, with the help of setting solder joints windows, the integrated projection method is developed to locate solder joints. Finally, the proposed location method is compared with other two algorithms in the experiment. Experiments result illustrates that the smaller the solder joint size, the better location accuracy and efficiency is obtained than other two mothers.

Lateral interactions across space reveal links between processing streams for luminance-modulated and contrast-modulated stimuli

Foveal detection thresholds for luminance-modulated (LM) and contrast-modulated (CM) blobs in the presence of fixed modulation, laterally placed noise blobs (separations of 0–6°) were measured in four observers with normal vision. Detection thresholds measured for LM blobs placed between highly visible LM flankers (1 1 1) and for CM blobs placed between highly visible CM flankers (2 2 2) produces a similar pattern of lateral interaction effects, i.e. masking where the stimuli overlap and facilitation for separations of 4–8× blob sd units. The region of facilitation is not matched by shallow psychometric function slopes. Detection thresholds measured for LM blobs placed between highly visible CM flankers (2 1 2) are generally facilitatory but relatively raised for separations of 0.5–2°. For CM blobs placed between highly visible LM flankers (1 2 1), facilitation is stronger in the 0.5–2° region. A significant correlation between thresholds and psychometric function slopes is found only for the 2 1 2 condition. We propose a model with two separate but interacting processing streams for the detection of LM and CM targets that may engage different cortical loci.

Fast and accurate extraction of moving object silhouette for personalized Virtual Reality Studio @ Home

Accurate segmentation of moving object silhouette in a real-time video is very important for object silhouette extraction in the vision-based interactive systems. However, the inherent problem of moving object segmentation based on the background subtraction criteria is to distinguish the changes occurring from background disturbing effects such as noise, shadows and illumination changes. The present paper proposes a hybrid method based on the background subtraction criteria that preserves the boundary of moving object and also robust against the noise and illumination changes. In the proposed method, the object regions are well identified by fusing the results from the background difference and motion-based change detection criterion. The shadows and highlights are well detected by utilizing the normalized luminance and background difference in Hue and Saturation component. The paper also introduces a novel connected component analysis procedure for detecting the object blob from the noise blobs, and a robust pixel-based background update scheme for updating the dynamic changes in the background. Moreover, the computational complexity of the proposed algorithm is analyzed. The proposed method has been implemented and evaluated regarding the segmentation quality and the frame rate. Further, the method has been shown to successfully extract the moving object silhouette and robust against the disturbing effects. Moreover, the proposed method has been tested in the VR@Home platform.

A new video segmentation method of moving objects based on blob-level knowledge

Variants of the background subtraction method are broadly used for the detection of moving objects in video sequences in different applications. In this work we propose a new approach to the background subtraction method which operates in the colour space and manages the colour information in the segmentation process to detect and eliminate noise. This new method is combined with blob-level knowledge associated with different types of blobs that may appear in the foreground. The idea is to process each pixel differently according to the category to which it belongs: real moving objects, shadows, ghosts, reflections, fluctuation or background noise. Thus, the foreground resulting from processing each image frame is refined selectively, applying at each instant the appropriate operator according to the type of noise blob we wish to eliminate. The approach proposed is adaptive, because it allows both the background model and threshold model to be updated. On the one hand, the results obtained confirm the robustness of the method proposed in a wide range of different sequences and, on the other hand, these results underline the importance of handling three colour components in the segmentation process rather than just the one grey-level component.