Microsoft Kinect sensor has been widely used in many applications since the launch of its first version. Recently, Microsoft released a new version of Kinect sensor with improved hardware. However, the accuracy assessment of the sensor remains to be answered. In this paper, we measure the depth accuracy of the newly released Kinect v2 depth sensor, and obtain a cone model to illustrate its accuracy distribution. We then evaluate the variance of the captured depth values by depth entropy. In addition, we propose a trilateration method to improve the depth accuracy with multiple Kinects simultaneously. The experimental results are provided to ascertain the proposed model and method.

Date fruits are small fruits that are abundant and popular in the Middle East, and have growing international presence. There are many different types of dates, each with different features. Sorting of dates is a key process in the date industry, and can be a tedious job. In this paper, we present a method for automatic classification of date fruits based on computer vision and pattern recognition. The method was implemented, and empirically tested on an image data spanning seven different categories of dates. In our method, an appropriately crafted mixture of fifteen different visual features was extracted, and then, multiple methods of classification were tried out, until satisfactory performance was achieved. Top accuracies ranged between 89% and 99%.

No abstract

People who suffer from hearing impairment caused by illness, age or extremely noisy environments are constantly in danger of being hit or knocked down by fast moving objects behind them when they have no companion or augmented sensory system to warn them. In this paper, we propose the General Moving Object Alarm System (GMOAS), a system focused on aiding the safe mobility of people under these circumstances. The GMOAS is a wearable haptic device that consists of two main subsystems: (i) a moving object monitoring subsystem that uses laser range data to detect and track approaching objects, and (ii) an alarm subsystem that warns the user of possibly dangerous approaching objects by triggering tactile vibrations on an ʺalarm necklaceʺ. For moving object monitoring, we propose a simple yet efficient solution to monitor the approaching behavior of objects. Compared with previous work in motion detection and tracking, we are not interested in specific objects but any type of approaching object that might harm the user. To this extent, we define a boundary in the laser range data where the objects are monitored. Within this boundary a fan‐shape grid is constructed to obtain an evenly distributed spatial partitioning of the data. These partitions are efficiently clustered into continuous objects which are then tracked through time using an object association algorithm based on updating a deviation matrix that represents angle, distance and size variations of the objects. The speed of the tracked objects is monitored throughout the algorithm. When the speed of an approaching object surpasses the safety threshold, the alarm necklace is triggered indicating the approaching direction of the fast moving object. The alarm necklace is equipped with three motors that can indicate five directions with respect to the user: left, back, right, left‐ back and right‐back. We performed three types of outdoor experiments (object passing, approaching and crossing) that empirically verified the effectiveness of our proposed algorithm. Furthermore, we analyzed the time and direction response based on neck vibrations. The statistical analysis (including hypothesis test) suggests that the chosen alarm necklace can provide a rapid indication for a quick human response

No abstract