In recent years, the accurate identification of chili maturity stages has become essential for optimizing cultivation processes. Conventional methodologies, primarily reliant on manual assessments or rudimentary detection systems, often fall short of reflecting the plant’s natural environment, leading to inefficiencies and prolonged harvest periods. Such methods may be imprecise and time-consuming. With the rise of computer vision and pattern recognition technologies, new opportunities in image recognition have emerged, offering solutions to these challenges. This research proposes an affordable solution for object detection and classification, specifically through version 5 of the You Only Look Once (YOLOv5) model, to determine the location and maturity state of rocoto chili peppers cultivated in Ecuador. To enhance the model’s efficacy, we introduce a novel dataset comprising images of chili peppers in their authentic states, spanning both immature and mature stages, all while preserving their natural settings and potential environmental impediments. This methodology ensures that the dataset closely replicates real-world conditions encountered by a detection system. Upon testing the model with this dataset, it achieved an accuracy of 99.99% for the classification task and an 84% accuracy rate for the detection of the crops. These promising outcomes highlight the model’s potential, indicating a game-changing technique for chili small-scale farmers, especially in Ecuador, with prospects for broader applications in agriculture. Doi: 10.28991/ESJ-2024-08-02-08 Full Text: PDF

This study sheds light on the evolution of the agricultural industry and highlights advances in production area. The salient recognition of fruit size and shape as critical quality parameters underscores the significance of the research. In response to this challenge, the research introduces specialized image processing techniques designed to streamline the sorting of apples in agricultural settings, specifically emphasizing accurate apple width estimation. A purpose-built machine was designed, featuring an enclosure box housing a cost-effective camera for the vision system and a chain conveyor for classifying Malus domestica Borkh kind apples. These goals were successfully achieved by implementing image preprocessing, segmentation, and measurement techniques to facilitate sorting. The proposed methodology classifies apples into three distinct classes, attaining an impressive accuracy of 94% in Class 1, 92% in Class 2, and 86% in Class 3. This represents an efficient and economical solution for apple classification and size estimation, promising substantial enhancements to sorting processes and pushing the boundaries of automation in the agricultural sector.

This report presents the development of a robotic arm for the efficient packaging of chocolates in a small-scale process. The robotic arm has a PRR (prismatic–rotational–rotational) configuration and uses a universal vacuum grip (UVG) to manipulate the chocolates with precision. Various materials and components were used, such as 3D-printed PLA parts for the mechanical elements and flexible TPU bands for optimal movement control. In terms of electronics, NEMA motors, an Arduino board, and a CNC interface were used for the precise control of the motors. The programming was conducted in Python and a graphical user interface (GUI) was created to manage the sequence of movements of the robotic arm. Key parameters, such as the accuracy in the placement of the chocolates, the grip strength, and the shape recovery ability were evaluated. The results demonstrate the successful integration of the robotic arm in the packaging process, achieving an efficient and consistent placement of the chocolates in the plastic trays. This project highlights the importance and potential of automation in the food packaging industry, by improving productivity, reducing human error, and ensuring packaging quality. The knowledge and results obtained in this project contribute to the field of robotics.

Eggs represent a fundamental food product worldwide due to their high protein content and many essential nutrients. The eggs’ delicate shells are susceptible to cracking, creating bacterial contamination opportunities and introducing reduced shelf life and monetary losses for agricultural business operations. Maintaining egg quality remains vital because it protects human safety and keeps production operations economical. Conventional quality control through manual methods has proven laborious and pruned to human mistakes while becoming unfeasible for industrial-scale production. This research develops an economical automated system that detects egg cracks while executing condition-based labeling and overcoming conventional inspection methods’ drawbacks. A machine vision system combines multiple computer vision methods that use color-based segmentation, morphological filtering, edge detection with the Canny method, and connected component analysis for digital surveys. The system classifies eggs into two classes: cracked and healthy; then, good condition eggs are labeled and sorted while cracked eggs are discarded. Experimental results achieved 94.89% classification accuracy and a 3-second from laying the egg to labeling, which provides potential industrial adoption. The proposed system offers benefits for reliable food safety measures and uses waste prevention features to help accelerate automated egg processing quality control. This system executes live analysis through a different approach. It analyzes eggs entering the testing area while they depart the post-harvest phase without necessitating any pre-existing datasets for training. The real-time functionality enables smooth operations in quality control procedures to support large-scale industrial use.

This paper elucidates the development of an advanced maturity detection system for cherry tomatoes, employing state-of-the-art computer vision technologies alongside automated irrigation systems. The proposed solution integrates sophisticated software with hardware components, incorporating moisture sensors and solenoid valves to ensure precise irrigation management. Central to this system is the YOLOv8 model, which has been specifically trained to differentiate between ripe and unripe tomatoes. A custom dataset comprising 1,442 images was meticulously curated and processed to facilitate the creation of the detection model utilizing the Roboflow platform. The training of the model was executed on Google Colab, capitalizing on its computational power to enhance model performance. Following comprehensive optimization, the YOLOv8 model attained an accuracy rate of 78.8%, establishing it as a reliable asset for agricultural applications. A user-friendly graphical interface, developed in Python with Tkinter, enables farmers to effectively monitor crop maturity and manage irrigation processes in real time. The validation results underscore the system’s remarkable efficacy in practical agricultural settings, emphasizing its potential to augment productivity and resource efficiency. This system signifies a considerable advancement in innovative agricultural technologies.

This work presents an intelligent feeding and hydration prototype for domestic cats that integrates deep-learning-based computer vision, embedded control, cloud communication, and a mobile interface. YOLO11n-seg and YOLO11s-seg were evaluated for cat detection, coat-region segmentation, and classification into solid, tabby, bicolour, and multicolour categories. YOLO11s-seg achieved the strongest bounding-box and mask precision and recall, while both models showed comparable detection performance in terms of mAP@50. Based on its superior segmentation results and acceptable computational requirements, YOLO11s-seg was selected for integration into the prototype. The validated visual output is transmitted to the control architecture and used as a contextual condition for food dispensing and water-management sequences. The ESP32 coordinates the stepper motor, presence and level sensors, pumps, and relays, while the mobile application enables parameter configuration, monitoring, and manual activation. Prototype testing confirmed the end-to-end operation of the vision, cloud, control, and actuation modules. The proposed system does not perform reliable individual-cat identification; instead, it demonstrates the feasibility of using lightweight coat-pattern segmentation and temporal validation to extend conventional timer-based pet-care devices.