SliceNet: A proficient model for real-time 3D shape-based recognition

Banana Pseudostem Width Detection Based on Kinect V2 Depth Sensor

This study used Kinect V2 sensor to collect the three-dimensional point cloud data of banana pseudostem and developed an automatic measurement method of banana pseudostem width. The banana plant was selected as the research object in a banana plantation in Fusui, Guangxi. The mobile measurement of banana pseudostem was carried out at a distance of 1 m from the banana plant using the field operation platform with Kinect V2 as the collection equipment. To eliminate the background data and improve the processing speed, a cascade classifier was used to recognize banana pseudostems from the depth image, extract the region of interest (ROI), and transform the ROI into a color point cloud combined with the color image; secondly, the point cloud was sparse by down-sampling; then, the point cloud noise was removed according to the classification of large-scale and small-scale noise; finally, the stem point cloud was segmented along the y-axis, and the difference between the maximum and minimum values in the x-axis direction of each segment was calculated as its horizontal width. The center point of each segment point cloud was used to fit the slope of the stem centerline, and the average horizontal width was corrected to the stem diameter. The test results show that the average measurement error is only 2.7 mm, the average relative error was 1.34%, and the measurement time is only about 300 ms. It could provide an effective solution for the automatic and rapid measurement of stem width of banana plants and other similar plants.

Improved calibration method of a four-quadrant detector based on Bayesian theory in a laser auto-collimation measurement system

Reliable and accurate calibration for a four-quadrant detector (QD) is a prerequisite for high-accuracy laser auto-collimation measurements. However, the calibration accuracy is limited largely by the non-linearity of QD, especially for large-range detection. To address this issue, an improved calibration method of QD based on Bayesian theory in laser auto-collimation measurement is proposed in this paper. First, the non-linearity problem of QD is analyzed, and for accurate calibration of QD, a high-precision identification model based on Bayesian theory is presented. An analytical expression between the output signal of QD and the position of the laser spot is established, and then a calibration system with laser drift compensation to avoid influences from the laser source as a datum is constructed. A series of experiments is conducted to verify the performance of the improved calibration method. The results reveal that the improved method can effectively enhance the calibration accuracy of QD and reduce the residuals in root mean square error by 86% compared to the 15-order polynomial fitting over a detection range of ±1mm. The comparison experiments also demonstrate that the proposed calibration method has advantages over the conventional method in terms of accuracy and robustness.

SparseVoxNet: 3-D Object Recognition With Sparsely Aggregation of 3-D Dense Blocks

Automatic recognition of 3-D objects in a 3-D model by convolutional neural network (CNN) methods has been successfully applied to various tasks, e.g., robotics and augmented reality. Three-dimensional object recognition is mainly performed by analyzing the object using multi-view images, depth images, graphs, or volumetric data. In some cases, using volumetric data provides the most promising results. However, existing recognition techniques on volumetric data have many drawbacks, such as losing object details on converting points to voxels and the large size of the input volume data that leads to substantial 3-D CNNs. Using point clouds could also provide very promising results; however, point-cloud-based methods typically need sparse data entry and time-consuming training stages. Thus, using volumetric could be a more efficient and flexible recognizer for our special case in the School of Medicine, Shanghai Jiao Tong University. In this article, we propose a novel solution to 3-D object recognition from volumetric data using a combination of three compact CNN models, low-cost SparseNet, and feature representation technique. We achieve an optimized network by estimating extra geometrical information comprising the surface normal and curvature into two separated neural networks. These two models provide supplementary information to each voxel data that consequently improve the results. The primary network model takes advantage of all the predicted features and uses these features in Random Forest (RF) for recognition purposes. Our method outperforms other methods in training speed in our experiments and provides an accurate result as good as the state-of-the-art.

Two-dimensional spectral signal model for chromatic confocal microscopy

In chromatic confocal microscopy, the signal characteristics influence the accuracy of the signal processing, which in turn determines measurement performance. Thus, a full understanding of the spectral characteristics is critical to enhance the measurement performance. Existing spectral models only describe the signal intensity-wavelength characteristics, without taking the displacement-wavelength relation into consideration. These models require prior knowledge of the optical design, which reduces the effectiveness in the optical design process. In this paper, we develop a two-dimensional spectral signal model to describe the signal intensity-wavelength-displacement characteristics in chromatic confocal microscopy without prior knowledge of the optical design layout. With this model, the influence of the dimensional characteristics of the confocal setup and the displacement-wavelength characteristics and monochromatic aberrations of the hyperchromatic objective are investigated. Experimental results are presented to illustrate the effectiveness of our signal model. Using our model, further evaluation of the spectral signal can be used to enhance the measurement performance of chromatic confocal microscopy.

Deep Learning Sensor Fusion for Autonomous Vehicle Perception and Localization: A Review

Autonomous vehicles (AV) are expected to improve, reshape, and revolutionize the future of ground transportation. It is anticipated that ordinary vehicles will one day be replaced with smart vehicles that are able to make decisions and perform driving tasks on their own. In order to achieve this objective, self-driving vehicles are equipped with sensors that are used to sense and perceive both their surroundings and the faraway environment, using further advances in communication technologies, such as 5G. In the meantime, local perception, as with human beings, will continue to be an effective means for controlling the vehicle at short range. In the other hand, extended perception allows for anticipation of distant events and produces smarter behavior to guide the vehicle to its destination while respecting a set of criteria (safety, energy management, traffic optimization, comfort). In spite of the remarkable advancements of sensor technologies in terms of their effectiveness and applicability for AV systems in recent years, sensors can still fail because of noise, ambient conditions, or manufacturing defects, among other factors; hence, it is not advisable to rely on a single sensor for any of the autonomous driving tasks. The practical solution is to incorporate multiple competitive and complementary sensors that work synergistically to overcome their individual shortcomings. This article provides a comprehensive review of the state-of-the-art methods utilized to improve the performance of AV systems in short-range or local vehicle environments. Specifically, it focuses on recent studies that use deep learning sensor fusion algorithms for perception, localization, and mapping. The article concludes by highlighting some of the current trends and possible future research directions.

Characterization of the displacement response in chromatic confocal microscopy with a hybrid radial basis function network

Characterization of the displacement response is critical for accurate chromatic confocal measurement. Current characterization methods usually provide a linear or polynomial relationship between the extracted peak wavelengths of the spectral signal and displacement. However, these methods are susceptible to errors in the peak extraction algorithms and errors in the selected model. In this paper, we propose a hybrid radial basis function network method to characterise the displacement response. With this method, the peak wavelength of the spectral signal is firstly extracted with a state-of-art peak extraction algorithm, following which, a higher-accuracy chromatic dispersion model is applied to determine the displacement-wavelength relationship. Lastly, a radial basis function network is optimized to provide a mapping between the spectral signals and the residual fitting errors of the chromatic dispersion model. Using experimental tests, we show that the hybrid radial basis function network method significantly improves the measurement accuracy, when compared to the existing characterizing methods.