scholarly journals A Real-Time Semantic Segmentation Method of Sheep Carcass Images Based on ICNet

2021 ◽  
Vol 2021 ◽  
pp. 1-12
Author(s):  
Shida Zhao ◽  
Guangzhao Hao ◽  
Yichi Zhang ◽  
Shucai Wang
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Processing Time ◽  
Image Data ◽  
Semantic Segmentation ◽  
Segmentation Method ◽  
Single Image ◽  
Evaluation Standard ◽  
Segmentation Accuracy ◽  
Image Dataset

How to realize the accurate recognition of 3 parts of sheep carcass is the key to the research of mutton cutting robots. The characteristics of each part of the sheep carcass are connected to each other and have similar features, which make it difficult to identify and detect, but with the development of image semantic segmentation technology based on deep learning, it is possible to explore this technology for real-time recognition of the 3 parts of the sheep carcass. Based on the ICNet, we propose a real-time semantic segmentation method for sheep carcass images. We first acquire images of the sheep carcass and use augmentation technology to expand the image data, after normalization, using LabelMe to annotate the image and build the sheep carcass image dataset. After that, we establish the ICNet model and train it with transfer learning. The segmentation accuracy, MIoU, and the average processing time of single image are then obtained and used as the evaluation standard of the segmentation effect. In addition, we verify the generalization ability of the ICNet for the sheep carcass image dataset by setting different brightness image segmentation experiments. Finally, the U-Net, DeepLabv3, PSPNet, and Fast-SCNN are introduced for comparative experiments to further verify the segmentation performance of the ICNet. The experimental results show that for the sheep carcass image datasets, the segmentation accuracy and MIoU of our method are 97.68% and 88.47%, respectively. The single image processing time is 83 ms. Besides, the MIoU of U-Net and DeepLabv3 is 0.22% and 0.03% higher than the ICNet, but the processing time of a single image is longer by 186 ms and 430 ms. Besides, compared with the PSPNet and Fast-SCNN, the MIoU of the ICNet model is increased by 1.25% and 4.49%, respectively. However, the processing time of a single image is shorter by 469 ms and expands by 7 ms, respectively.

scholarly journals A Study on Semantic Segmentation for Autonomous Vehicles

2020 ◽  
Author(s):  
Vinícius Almeida dos Santos ◽  
Rodrigo Lyra ◽  
Thiago Felski Pereira
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Autonomous Vehicles ◽  
Processing Time ◽  
Semantic Segmentation ◽  
Time Constraint ◽  
Segmentation Algorithm ◽  
Digital Cameras ◽  
Important Challenge

Autonomous vehicles are already a reality, and there are still severalchallenges to overcome. One important challenge for the adoptionof these vehicles is perceiving its surroundings. This necessity ofperception can be fulfilled by digital cameras. When working withdigital image processing, the quality will be limited by real-timeconstraints. As several works indicate, this real-time constraint forautonomous vehicles is at most 100ms per frame. Also, by improvingthe processing time, the chances of accidents involving autonomousvehicles may be decreased. This paper analyses the advantages anddrawbacks of semantic segmentation and also presents a study toimplement perception for autonomous vehicles by accelerating asemantic segmentation algorithm, also used by other works on thefield. To accelerate the algorithm, spacial parallelism will be used.

scholarly journals InstantDL: an easy-to-use deep learning pipeline for image segmentation and classification

2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Dominik Jens Elias Waibel ◽  
Sayedali Shetab Boushehri ◽  
Carsten Marr
Keyword(s):  
Image Processing ◽  
Deep Learning ◽  
Specific Problem ◽  
State Of The Art ◽  
Image Data ◽  
Semantic Segmentation ◽  
Parameter Tuning ◽  
Cellular Processes ◽  
Minimal Effort ◽  
Instance Segmentation

Abstract Background Deep learning contributes to uncovering molecular and cellular processes with highly performant algorithms. Convolutional neural networks have become the state-of-the-art tool to provide accurate and fast image data processing. However, published algorithms mostly solve only one specific problem and they typically require a considerable coding effort and machine learning background for their application. Results We have thus developed InstantDL, a deep learning pipeline for four common image processing tasks: semantic segmentation, instance segmentation, pixel-wise regression and classification. InstantDL enables researchers with a basic computational background to apply debugged and benchmarked state-of-the-art deep learning algorithms to their own data with minimal effort. To make the pipeline robust, we have automated and standardized workflows and extensively tested it in different scenarios. Moreover, it allows assessing the uncertainty of predictions. We have benchmarked InstantDL on seven publicly available datasets achieving competitive performance without any parameter tuning. For customization of the pipeline to specific tasks, all code is easily accessible and well documented. Conclusions With InstantDL, we hope to empower biomedical researchers to conduct reproducible image processing with a convenient and easy-to-use pipeline.

scholarly journals Lightweight Semantic Segmentation Network for Real-Time Weed Mapping Using Unmanned Aerial Vehicles

2020 ◽  
Vol 10 (20) ◽  
pp. 7132 ◽  
Author(s):  
Jizhong Deng ◽  
Zhaoji Zhong ◽  
Huasheng Huang ◽  
Yubin Lan ◽  
Yuxing Han ◽  
...  
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Network Architecture ◽  
Semantic Segmentation ◽  
Real Time Image Processing ◽  
Weed Mapping ◽  
Real Time Image ◽  
Time Gap ◽  
Image Collection ◽  
Time Image

The timely and efficient generation of weed maps is essential for weed control tasks and precise spraying applications. Based on the general concept of site-specific weed management (SSWM), many researchers have used unmanned aerial vehicle (UAV) remote sensing technology to monitor weed distributions, which can provide decision support information for precision spraying. However, image processing is mainly conducted offline, as the time gap between image collection and spraying significantly limits the applications of SSWM. In this study, we conducted real-time image processing onboard a UAV to reduce the time gap between image collection and herbicide treatment. First, we established a hardware environment for real-time image processing that integrates map visualization, flight control, image collection, and real-time image processing onboard a UAV based on secondary development. Second, we exploited the proposed model design to develop a lightweight network architecture for weed mapping tasks. The proposed network architecture was evaluated and compared with mainstream semantic segmentation models. Results demonstrate that the proposed network outperform contemporary networks in terms of efficiency with competitive accuracy. We also conducted optimization during the inference process. Precision calibration was applied to both the desktop and embedded devices and the precision was reduced from FP32 to FP16. Experimental results demonstrate that this precision calibration further improves inference speed while maintaining reasonable accuracy. Our modified network architecture achieved an accuracy of 80.9% on the testing samples and its inference speed was 4.5 fps on a Jetson TX2 module (Nvidia Corporation, Santa Clara, CA, USA), which demonstrates its potential for practical agricultural monitoring and precise spraying applications.

Design and Realization of Double-Buses Modularized Structure Image Processing System

2011 ◽  
Vol 179-180 ◽  
pp. 257-263
Author(s):  
Biao Zhang ◽  
Yue Huan Wang
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Processing System ◽  
Image Data ◽  
Parallel Structure ◽  
Image Processing System ◽  
Real Time Image Processing ◽  
Real Time Image ◽  
Data Bus ◽  
Time Image

It is double-buses modularized structure with the combination of system control bus and high speed image data bus which is put forward in this paper. Moreover, the management and distribution of image data bus and the design of system reset procedure are elaborated through which a kind of practical real-time image processing system with the strongest adaptability and capability for structure programming and system expansion. The computing capability in infrared test of small target is greatly improved which is verified in tri DSP model system. According to complex image processing task, through the adjustment of parallel structure of image processing algorithm, the higher parallel efficiency can be realized. So to say, the system structure has a great adjustment to algorithm parallel structure and can be successfully used as a platform for universal real-time image processing.

scholarly journals Segmentation Comparing Eggs Watermarking Image and Original Image

2017 ◽  
Vol 6 (1) ◽  
pp. 47-53
Author(s):  
Anton Yudhana ◽  
Sunardi Sunardi ◽  
Shoffan Saifullah
Keyword(s):  
Image Processing ◽  
Small Area ◽  
Image Data ◽  
Watermark Image ◽  
Original Image ◽  
Segmentation Method ◽  
Bounding Box ◽  
Chicken Eggs ◽  
The Impact

The research used watermarking techniques to obtain the image originality. The aims of the research were to identify small area in eggs properly and compared preprocessing, the methods, and the results of image processing. The study has been improved from the previous papers by combined all methods and analysis was obtained.This study was conducted by using centroid and the bounding box for determining the object and the small area of chicken eggs. The segmentation method was used to compare the original image and the watermarked image. Image processing using image data that are subject watermark to maintain the authenticity of the images used in the study will the impact in delivering the desired results. In the identification of chicken eggs using watermark image using several methods are expected to provide results as desired. Segmentation also deployed to process the Image and counted the objects. The results showed that the process of segmentation and objects counting determined that the original image and watermarked image had the same value and recognized eggs. Identification had determined percentage of 100% for all the samples.

Techniques for the Automated Segmentation of Lung in Thoracic Computed Tomography Scans

2013 ◽  
pp. 675-687
Author(s):  
William F. Sensakovic ◽  
Samuel G. Armato
Keyword(s):  
Computed Tomography ◽  
Image Data ◽  
Segmentation Method ◽  
Correlation Filters ◽  
Lung Segmentation ◽  
Thoracic Diseases ◽  
Thoracic Computed Tomography ◽  
Segmentation Accuracy ◽  
Computed Tomography Scans ◽  
Computer Aided

Computed Tomography (CT) is widely used to diagnose and assess thoracic diseases. The improved resolution of CT studies has resulted in a substantial increase of image data for analysis by radiologists. The time-consuming nature of this analysis motivates the application of Computer-Aided Diagnostic (CAD) methods to assist radiologists. Most CAD methods require identification of the lung within the patient images, a preprocessing step known as “lung segmentation.” This chapter describes an intensity-based lung segmentation method. The segmentation method begins with simple thresholding, and several image processing modules are included to improve segmentation accuracy and robustness. Common segmentation difficulties are discussed and motivate the inclusion of each module in the lung segmentation method. These modules will include brief explanations of common techniques (e.g., morphological operators) in addition to novel techniques developed specifically for lung segmentation (e.g., gradient correlation filters).

scholarly journals Image Processing for Sustainable Remodeling: Introduction to Real-time Quality Inspection System of External Wall Insulation Works

2019 ◽  
Vol 11 (4) ◽  
pp. 1081 ◽  
Author(s):  
Sang-Ho Cho ◽  
Kyung-Tae Lee ◽  
Se-Heon Kim ◽  
Ju-Hyung Kim
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Image Data ◽  
Inspection System ◽  
Quality Inspection ◽  
Labor Costs ◽  
External Wall ◽  
Graphic Data ◽  
Repair Costs ◽  
Serious Injury

The external wall insulation method was introduced to enhance the energy efficiency of existing buildings. It does not cause a decrease of inner space and costs less in comparison to methods that insert insulation panels inside walls. However, it has been reported that external wall insulation boards are disconnecting from walls due to malfunctions of the adhesive. This causes not only repair costs, but also serious injury to pedestrians. Separation problems occur when the bonded positions are incorrect and/or the total area and thickness of the adhesive is smaller than the required amount. A challenge is that these faults can hardly be inspected after installing boards. For this reason, a real-time inspection system is necessary to detect potential failure during adhesive works. Position, area and thickness are major aspects to inspect, and thus a method to process image data of these seems efficient. This paper presents a real-time quality inspection system introducing image processing technology to detect potential errors during adhesive works of external wall insulation, and it is predicted to contribute to achieving sustainable remodeling construction by reducing squandered material and labor costs. The system consists of a graphic data creation module to capture the results of adhesive works and a quality inspection module to judge the pass or fail of works according to an algorithm. A prototype is developed and validated against 100 panels with 800 adhesive points.

scholarly journals Embedded Real-Time Image Processing System Based on DM6446+FPGA Architecture

2012 ◽  
Vol 6-7 ◽  
pp. 542-546
Author(s):  
Bao Feng Zhang ◽  
Yi Yang ◽  
Jun Chao Zhu ◽  
Cui Li
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Processing System ◽  
Image Data ◽  
Image Processing System ◽  
Real Time Image Processing ◽  
Fpga Architecture ◽  
Real Time Image ◽  
Output Field ◽  
Time Image

To solve the traditional image processing system problem such as large in size, high power consumption and poor real-time, an embedded real-time image processing system is designed based on TMS320DM6446+FPGA architecture. DM6446 as the core of the system is responsible for the scheduling, image processing algorithms, image output; field programmable gate array (FPGA) is responsible for capturing real-time image data, image preprocessing. The paper describes the principle of the real-time image processing system. The experiment proved that the system can achieve real-time acquisition, processing and output of image data in 20 frames per second.

Research on ROI Image Processing Technology of Teleoperation Construction Robot Based on Trinocular Stereo Vision

2011 ◽  
Vol 130-134 ◽  
pp. 2581-2584
Author(s):  
Ming De Gong ◽  
Bo Tian ◽  
Yue Ning ◽  
Wei Wei Li
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Three Dimensional ◽  
Image Data ◽  
Robot System ◽  
The Real ◽  
Human Eye ◽  
Long Time ◽  
Calibration Algorithm

Digital image has a large quantity of image data and long time for transmitting. It affects the real-time of the teleoperation robot system. According to the basic principle of human eye identifying objects and image blurry processing, a new image processing method of simulating human eye range of interest (ROI) is proposed. The method uses the calibration algorithm of three-dimensional stereo target and the Gauss blurred principle. The non-ROI region is blurred to hierarchy for extracting the feature and measurement to finish the image processing tasks. The experimental results showed that the quality of the images was assured and the transmission time was shorted. The real-time of the teleoperation robot system was also guaranteed.

The System of Real-Time Image Processing and Combustion Diagnosis Based on Omap3530

2013 ◽  
Vol 816-817 ◽  
pp. 535-539
Author(s):  
Yu Jie Zhang ◽  
Ying Ying Wu
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Combustion Process ◽  
Engineering Application ◽  
Image Data ◽  
Signal Acquisition ◽  
Real Time Image Processing ◽  
Real Time Image ◽  
And Control ◽  
Time Image

In this paper, according to the characteristics of power plant boiler combustion process, to use the image processing technology to extract feature quantity of flame combustion. For boiler combustion diagnosis real-time requirements, the system of real-time image processing and combustion diagnosis based on OMAP3530 was designed and developed. The system makes full use of the OMAP3530 dual-core processor, and makes the operating system and control, video signal acquisition, human-computer interaction, output driving tasks run on the ARM, and the image data processing tasks are completed by DSP. It maximizes the performance of OMAP3530, improves the real-time performance of the system. Experiments were carried out in 200MW boiler. The results show that, the system is simple and practical, which can realize the combustion diagnosis of the running boiler and provide the reliable basis for the safety and economic operation of power station boiler. It has a certain engineering application prospect.

Sign in / Sign up

Export Citation Format

Share Document