A Wearable Mobile Sensor Platform to Assist Fruit Grading

Aroca, Rafael Vidal; Gomes, Rafael Beserra; Dantas, Rummenigge Rudson; Calbo, Adonai Gimenez; Gonçalves, Luiz Marcos Garcia

doi:10.3390/s130506109

Cited by 14 publications

(10 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, the inclusion of attention in the earlier steps of deep learning is planned to be implemented in a dedicated architecture with the GPU (Xavier from nVidia) to speed up processing in these kinds of devices. This future implementation is planned to work as the vision-processing system embarked in our wearable glove device [28], and also in payload applications in our robotic platforms, as drones and a robotic sailboat [29].…”

Section: Resultsmentioning

confidence: 99%

“…As an immediate application, this implementation will work as the vision-processing system embarked on our wearable glove device [28], and also in decision-making algorithms for payload applications in our robotic platforms, as drones, and in a robotic sailboat [29]. Applications already being developed for such platforms are such as visual attention and objects/target recognition for harvesting (the glove), navigation, and localization (in the sailboat).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dynamic multifoveated structure for real-time vision tasks in robotic systems

Medeiros

Gomes

Clua

et al. 2019

J Real-Time Image Proc

Self Cite

View full text Add to dashboard Cite

Foveation is a technique that allows real-time image processing by drastically reducing the amount of visual data without loosing essential information around some focused area. When a robot needs to pay attention at two or more regions of the image at the same time, e.g., for tracking two or more objects, multifoveation is necessary. In this case, computing features twice in the intersections between the different foveated structures, which could linearly increase the processing time, must be avoided. To solve this redundancy removal problem, we propose two algorithms. The first one is based on the previous calculation of redundant blocks and the second one is based on a pixel-by-pixel processing at execution time. Experimental results show a gain in processing time for the block-based model in comparison with the pixel-by-pixel and also of both in comparison with other approaches that sequentially calculate various single foveated images. Robotics vision and other tasks related to dynamic visual attention, as recognition, real-time surveillance, video transmission, and image rendering, are examples of applications that can rely on and strongly benefit from such model.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Dynamic multifoveated structure for real-time vision tasks in robotic systems

Medeiros

Gomes

Clua

et al. 2019

J Real-Time Image Proc

Self Cite

View full text Add to dashboard Cite

show abstract

“…Also, two applications (in simple depth maps correction and point clouds registration) are devised and implemented just to illustrate the utility of the depth RMS errors resulting from our approach. Other applications as wearable sensors [ 6 ] or probabilistic robotics [ 7 , 8 ] can also benefit from the RMS error estimation provided by our approach. In this last specific application, previously determined error bounding in function of the distance to points in the environment are used in order to estimate the variance in the depth provided by stereo cameras, which is used for mapping and reconstruction.…”

Section: Introductionmentioning

confidence: 99%

A Versatile Method for Depth Data Error Estimation in RGB-D Sensors

Cabrera

Ortiz

Silva

et al. 2018

Sensors

Self Cite

View full text Add to dashboard Cite

We propose a versatile method for estimating the RMS error of depth data provided by generic 3D sensors with the capability of generating RGB and depth (D) data of the scene, i.e., the ones based on techniques such as structured light, time of flight and stereo. A common checkerboard is used, the corners are detected and two point clouds are created, one with the real coordinates of the pattern corners and one with the corner coordinates given by the device. After a registration of these two clouds, the RMS error is computed. Then, using curve fittings methods, an equation is obtained that generalizes the RMS error as a function of the distance between the sensor and the checkerboard pattern. The depth errors estimated by our method are compared to those estimated by state-of-the-art approaches, validating its accuracy and utility. This method can be used to rapidly estimate the quality of RGB-D sensors, facilitating robotics applications as SLAM and object recognition.

show abstract

“…In recent years, with the increasing popularity of smart/wearable mobile sensor devices [ 10 ], such as smart phones, smart watches and smart glasses, various real-time applications, such as image code recognition, face recognition, optical character recognition (OCR), logo recognition, augmented reality (AR) and mixed reality (MR), have emerged for sensor computing [ 11 , 12 , 13 , 14 , 15 ]. Because smart/wearable mobile sensor devices possess limited hardware resources, such as low-performance processors, small-sized memory, low-resolution displays and low battery capacity, they require simple and fast algorithms for image processing.…”

Section: Introductionmentioning

confidence: 99%

Fast Contour-Tracing Algorithm Based on a Pixel-Following Method for Image Sensors

Seo¹,

Chae

Shim

et al. 2016

Sensors

View full text Add to dashboard Cite

Contour pixels distinguish objects from the background. Tracing and extracting contour pixels are widely used for smart/wearable image sensor devices, because these are simple and useful for detecting objects. In this paper, we present a novel contour-tracing algorithm for fast and accurate contour following. The proposed algorithm classifies the type of contour pixel, based on its local pattern. Then, it traces the next contour using the previous pixel’s type. Therefore, it can classify the type of contour pixels as a straight line, inner corner, outer corner and inner-outer corner, and it can extract pixels of a specific contour type. Moreover, it can trace contour pixels rapidly because it can determine the local minimal path using the contour case. In addition, the proposed algorithm is capable of the compressing data of contour pixels using the representative points and inner-outer corner points, and it can accurately restore the contour image from the data. To compare the performance of the proposed algorithm to that of conventional techniques, we measure their processing time and accuracy. In the experimental results, the proposed algorithm shows better performance compared to the others. Furthermore, it can provide the compressed data of contour pixels and restore them accurately, including the inner-outer corner, which cannot be restored using conventional algorithms.

show abstract

A Wearable Mobile Sensor Platform to Assist Fruit Grading

Cited by 14 publications

References 46 publications

Dynamic multifoveated structure for real-time vision tasks in robotic systems

Dynamic multifoveated structure for real-time vision tasks in robotic systems

A Versatile Method for Depth Data Error Estimation in RGB-D Sensors

Fast Contour-Tracing Algorithm Based on a Pixel-Following Method for Image Sensors

Contact Info

Product

Resources

About