Apple dezvolta de ceva vreme un sistem de ghidare autonoma pentru masini, acesta fiind gandit initial pentru o masina produsa de catre compania americana. Avand in vedere ca acel proiect a fost anulat de catre compania Apple, cei din Cupertino continua cu unul care are la baza dezvoltarea unui sistem de ghidare autonoma pentru masini.
Acest sistem ar urma sa fie vandut de catre Apple catre producatorii de automobile, iar astazi aflam cateva informatii interesante in legatura cu el. Apple se bazeaza pe tehnologia LIDAR pentru a crea acest sistem, el fiind capabil sa recunoasca pietoni, masini si ciclisti de la distante mari, o serie de camere si un software special fiind folosite pentru acest proces descris aici.
“Accurate detection of objects in 3D point clouds is a central problem in many applications, such as autonomous navigation, housekeeping robots, and augmented/virtual reality. To interface a highly sparse LiDAR point cloud with a region proposal network (RPN), most existing efforts have focused on hand-crafted feature representations, for example, a bird’s eye view projection. In this work, we remove the need of manual feature engineering for 3D point clouds and propose VoxelNet, a generic 3D detection network that unifies feature extraction and bounding box prediction into a single stage, end-to-end trainable deep network.”
Apple sustine ca tehnologia dezvoltata de ea poate recunoaste oameni si ciclisti mai bine decat alte tehnologii similare care au menirea de a detecta obiecte 3D. Deocamdata Apple a facut doar experimente software, fara a trece la teste reale, asa ca tehnologia nu este pregatita in totalitate pentru a fi implementata in vreun automobil.
Pana in momentul in care s-ar putea intampla asta, cei de la Apple le permit inginerilor sa publice lucrari care au legatura cu tehnologia pe care o dezvolta. Scopul acestui proces este de a aduce informatiile la cat mai multe persoane pentru ca ei sa obtina ajutor in dezvoltarea software-ului si a hardware-ului, dar ramane de vazut cat succes vor avea.
“Most existing methods in LiDAR-based 3D detection rely on hand-crafted feature representations, for example, a bird’s eye view projection. In this paper, we remove the bottleneck of manual feature engineering and propose Vox- elNet, a novel end-to-end trainable deep architecture for point cloud based 3D detection. Our approach can operate directly on sparse 3D points and capture 3D shape information effectively. We also present an efficient implementation of VoxelNet that benefits from point cloud sparsity and parallel processing on a voxel grid.”
Compania Apple este foarte hotarata sa aduca acest sistem de ghidare autonoma pe piata, el este asteptat de foarte multa lume si va fi extrem de interesant de vazut daca va avea succes, sau nu.
“Our experiments on the KITTI car detection task show that VoxelNet outper- forms state-of-the-art LiDAR based 3D detection methods by a large margin. On more challenging tasks, such as 3D detection of pedestrians and cyclists, VoxelNet also demonstrates encouraging results showing that it provides a better 3D representation”.
