Temporal changes in Lycium barbarum fruit separation force and hardness during selective harvesting.

Lycium barbarum, a plant belonging to the Solanaceae family, is widely used in China due to its abundant nutritional value. Although the current mechanized harvesting method of L. barbarum has effectively minimized production expenses, it continues to have the challenge of inconsistent quality of the produced L. barbarum. The objective of this paper is to evaluate the correlation of the separating force and hardness concerning the timing of harvesting, maturity, and variety. Thus, the optimal time for harvesting ripe L. barbarum can be determined to enhance the quality of selectively mechanized harvesting of this fruit. The experiment was conducted in a L. barbarum plantation located in Qinghai Province during the 2023 harvest period. Two occasions were studied focusing on the primary cultivars Ningqi No. 1 and Ningqi No. 7, examining the three ripening stages of L. barbarum harvested at various times throughout the day. The finding of this study showed that the separation force and hardness of L. barbarum fruits were influenced by the harvesting time, the fruit variety, and the level of maturity. The optimal timing for harvesting different types of L. barbarum varies. It was observed that Ningqi No.1 was best to be harvested in the late afternoon and evening (17:00-21:00), whereas Ningqi No.7 was most suitable to be harvested in the morning (7:00-9:00).

LiDAR and Deep Learning-Based Standing Tree Detection for Firebreaks Applications.

Forest fire prevention is very important for the protection of the ecological environment, which requires effective prevention and timely suppression. The opening of the firebreaks barrier contributes significantly to forest fire prevention. The development of an artificial intelligence algorithm makes it possible for an intelligent belt opener to create the opening of the firebreak barrier. This paper introduces an innovative vision system of an intelligent belt opener to monitor the environment during the creation of the opening of the firebreak barrier. It can provide precise geometric and location information on trees through the combination of LIDAR data and deep learning methods. Four deep learning networks including PointRCNN, PointPillars, SECOND, and PV-RCNN were investigated in this paper, and we train each of the four networks using our stand tree detection dataset which is built on the KITTI point cloud dataset. Among them, the PointRCNN showed the highest detection accuracy followed by PV-RCNN and PV-RCNN. SECOND showed less detection accuracy but can detect the most targets.

Determining the Stir-Frying Degree of Gardeniae Fructus Praeparatus Based on Deep Learning and Transfer Learning.

Gardeniae Fructus (GF) is one of the most widely used traditional Chinese medicines (TCMs). Its processed product, Gardeniae Fructus Praeparatus (GFP), is often used as medicine; hence, there is an urgent need to determine the stir-frying degree of GFP. In this paper, we propose a deep learning method based on transfer learning to determine the stir-frying degree of GFP. We collected images of GFP samples with different stir-frying degrees and constructed a dataset containing 9224 images. Five neural networks were trained, including VGG16, GoogLeNet, Resnet34, MobileNetV2, and MobileNetV3. While the model weights from ImageNet were used as initial parameters of the network, fine-tuning was used for four neural networks other than MobileNetV3. In the training of MobileNetV3, both feature transfer and fine-tuning were adopted. The accuracy of all five models reached more than 95.82% in the test dataset, among which MobileNetV3 performed the best with an accuracy of 98.77%. In addition, the results also showed that fine-tuning was better than feature transfer in the training of MobileNetV3. Therefore, we conclude that deep learning can effectively recognize the stir-frying degree of GFP.

Few-Shot Fine-Grained Forest Fire Smoke Recognition Based on Metric Learning.

To date, most existing forest fire smoke detection methods rely on coarse-grained identification, which only distinguishes between smoke and non-smoke. Thus, non-fire smoke and fire smoke are treated the same in these methods, resulting in false alarms within the smoke classes. The fine-grained identification of smoke which can identify differences between non-fire and fire smoke is of great significance for accurate forest fire monitoring; however, it requires a large database. In this paper, for the first time, we combine fine-grained smoke recognition with the few-shot technique using metric learning to identify fire smoke with the limited available database. The experimental comparison and analysis show that the new method developed has good performance in the structure of the feature extraction network and the training method, with an accuracy of 93.75% for fire smoke identification.

Full-Scale Fire Smoke Root Detection Based on Connected Particles.

Smoke is an early visual phenomenon of forest fires, and the timely detection of smoke is of great significance for early warning systems. However, most existing smoke detection algorithms have varying levels of accuracy over different distances. This paper proposes a new smoke root detection algorithm that integrates the static and dynamic features of smoke and detects the final smoke root based on clustering and the circumcircle. Compared with the existing methods, the newly developed method has a higher accuracy and detection efficiency on the full scale, indicating that the method has a wider range of applications in the quicker detection of smoke in forests and the prevention of potential forest fire spread.

An Embedded Portable Lightweight Platform for Real-Time Early Smoke Detection.

The advances in developing more accurate and fast smoke detection algorithms increase the need for computation in smoke detection, which demands the involvement of personal computers or workstations. Better detection results require a more complex network structure of the smoke detection algorithms and higher hardware configuration, which disqualify them as lightweight portable smoke detection for high detection efficiency. To solve this challenge, this paper designs a lightweight portable remote smoke front-end perception platform based on the Raspberry Pi under Linux operating system. The platform has four modules including a source video input module, a target detection module, a display module, and an alarm module. The training images from the public data sets will be used to train a cascade classifier characterized by Local Binary Pattern (LBP) using the Adaboost algorithm in OpenCV. Then the classifier will be used to detect the smoke target in the following video stream and the detected results will be dynamically displayed in the display module in real-time. If smoke is detected, warning messages will be sent to users by the alarm module in the platform for real-time monitoring and warning on the scene. Case studies showed that the developed system platform has strong robustness under the test datasets with high detection accuracy. As the designed platform is portable without the involvement of a personal computer and can efficiently detect smoke in real-time, it provides a potential affordable lightweight smoke detection option for forest fire monitoring in practice.

Homogenized modeling methodology for 18650 lithium-ion battery module under large deformation.

Effective lithium-ion battery module modeling has become a bottleneck for full-size electric vehicle crash safety numerical simulation. Modeling every single cell in detail would be costly. However, computational accuracy could be lost if the module is modeled by using a simple bulk material or rigid body. To solve this critical engineering problem, a general method to establish a computational homogenized model for the cylindrical battery module is proposed. A single battery cell model is developed and validated through radial compression and bending experiments. To analyze the homogenized mechanical properties of the module, a representative unit cell (RUC) is extracted with the periodic boundary condition applied on it. An elastic-plastic constitutive model is established to describe the computational homogenized model for the module. Two typical packing modes, i.e., cubic dense packing and hexagonal packing for the homogenized equivalent battery module (EBM) model, are targeted for validation compression tests, as well as the models with detailed single cell description. Further, the homogenized EBM model is confirmed to agree reasonably well with the detailed battery module (DBM) model for different packing modes with a length scale of up to 15 × 15 cells and 12% deformation where the short circuit takes place. The suggested homogenized model for battery module makes way for battery module and pack safety evaluation for full-size electric vehicle crashworthiness analysis.