State of Art IoT and Edge Embedded Systems for Real-Time Machine Vision Applications

Meribout, Mahmoud; Baobaid, Asma; Khaoua, Mohammed Ould; Pena, Juan Pablo

doi:10.1109/access.2022.3175496

Cited by 16 publications

(6 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Image analysis that is carried out within embedded systems for shape classification and recognition should consider their limited computational power and reduced mem-ory [77,78]. Memory size will depend on the number of arrays per image [79,80], whether they are RGB or grayscale, and their quality (bits per pixel).…”

Section: Discussionmentioning

confidence: 99%

Mango Fruit Fly Trap Detection Using Different Wireless Communications

Hahn,

Valle,

Rendón

et al. 2023

Agronomy

View full text Add to dashboard Cite

Fruit flies cause production losses in mango orchards affecting fruit quality. A National Campaign against Fruit Flies (NCFF) evaluates farm status using the fruit flies per trap per day index (FTD). Traps with attractant are installed manually within orchards in Mexico, but counting the flies trapped every week requires excessive numbers of trained personal. Electronic traps (e-traps) use sensors to monitor fruit fly population, saving labor and obtaining the real-time orchard infestation. The objective of this work was to acquire an image within a e-trap at 17:00 when an insect was detected and binarize the information in real-time to count the number of flies. Each e-trap was implemented with a polyethylene PET bottle screwed to a tap containing an ESP32-CAM camera. E-traps from several hectares of mango trees were sampled and transmitted through WSN wireless sensor networks. This original system presents a star topology network within each hectare with the long range LoRa transceiver at the central tower. It receives the fly count from five e-traps and finally transmits data to the house tower end point. Another contribution of this research was the use of a DJI mini2 for acquiring the e-trap data, and the 8-ha flight took 15 min and 35 s. This period can be reduced if the drone flies higher.

show abstract

Section: Discussionmentioning

confidence: 99%

Mango Fruit Fly Trap Detection Using Different Wireless Communications

Hahn,

Valle,

Rendón

et al. 2023

Agronomy

View full text Add to dashboard Cite

show abstract

“…However, for simple, low-processing applications, using low-power and cheap devices is of interest [33,34]. These devices can be off-the-shelf single-board computers [35] or custom-built embedded systems [2,35,36].…”

Section: Related Workmentioning

confidence: 99%

Performance Evaluation of Computer Vision Algorithms in a Programmable Logic Controller: An Industrial Case Study

Vieira,

Silva,

Ribeiro

et al. 2024

Sensors

View full text Add to dashboard Cite

This work evaluates the use of a programmable logic controller (PLC) from Phoenix Contact’s PLCnext ecosystem as an image processing platform. PLCnext controllers provide the functions of “classical” industrial controllers, but they are based on the Linux operating system, also allowing for the use of software tools usually associated with computers. Visual processing applications in the Python programming language using the OpenCV library are implemented in the PLC using this feature. This research is focused on evaluating the use of this PLC as an image processing platform, particularly for industrial machine vision applications. The methodology is based on comparing the PLC’s performance against a computer using standard image processing algorithms. In addition, a demonstration application based on a real-world scenario for quality control by visual inspection is presented. It is concluded that despite significant limitations in processing power, the simultaneous use of the PLC as an industrial controller and image processing platform is feasible for applications of low complexity and undemanding cycle times, providing valuable insights and benchmarks for the scientific community interested in the convergence of industrial automation and computer vision technologies.

show abstract

“…In contrast to the positive kinematic analysis of the arm, the inverse kinematic analysis of the arm uses calculations to solve for the angle of rotation at each joint when the relevant parameters of the connecting rod and the position of the end of the arm concerning the reference coordinate system have been obtained [23]. Geometric analysis based on the structure of the robotic arm: for the robotic arm involved in this paper, it is necessary to find the angle of rotation at each joint when the end of the robotic arm is moved from O to O , i.e., to find γ, α, and β, according to the inverse kinematics requirements.…”

Section: Robotic Arm Inverse Kinematics Analysismentioning

confidence: 99%

Design of a Four-Axis Robot Arm System Based on Machine Vision

Wang,

Zhou,

Wei

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

With the concept of industrial automation gradually being put forward, the four-axis robotic arm is gradually being applied in industrial production environments due to its advantages such as a stable structure, easy maintenance, and expandability. However, it is difficult to diversify and improve the traditional four-axis robotic arm system due to the high software and hardware coupling and the single system design, which results in high production costs. At the same time, its low intelligence and high-power consumption limit its wide application. The paper proposes an embedded design of a four-axis manipulator system based on vision guidance. Based on the robot kinematics theory and geometric principles, the dynamics simulation of the manipulator model is carried out. Through the forward and reverse analysis of the manipulator model and the trajectory planning of the manipulator, the YOLOV7 target detection algorithm is introduced and deployed on the embedded device, which greatly reduces the manufacturing cost of the manipulator while meeting the control and power consumption requirements. It has been verified by experiments that the robot arm in this paper can achieve an end accuracy of 0.05 mm under the condition of a load of 1 kg using the ISO 9283 international standard, and the recognition algorithm adopted can achieve a recognition accuracy of 95.2% at a frame rate of 29. The overall power consumption is also lower than that of traditional robotic arms.

show abstract

State of Art IoT and Edge Embedded Systems for Real-Time Machine Vision Applications

Cited by 16 publications

References 33 publications

Mango Fruit Fly Trap Detection Using Different Wireless Communications

Mango Fruit Fly Trap Detection Using Different Wireless Communications

Performance Evaluation of Computer Vision Algorithms in a Programmable Logic Controller: An Industrial Case Study

Design of a Four-Axis Robot Arm System Based on Machine Vision

Contact Info

Product

Resources

About