(Original title: Tesla ’s Neural Net can detect traffic light colors, garbage cans, and detailed road markings)

Netease Technology News on December 25, according to foreign media reports, some Tesla owners have received a preview version of the "FSD (Full Automated Driving) Kit", showing that the electric vehicle manufacturer's neural network has been greatly improved. More common objects can be detected and identified, such as traffic lights, trash cans, and lane markings.

Tesla owner and enthusiast Steve Hamel, after receiving the latest 2019.40.50 software update, recorded some of the new features of the Model 3 and shared several photos.

The photos show that the neural network on Tesla's car is now able to recognize trash cans, lane lines, arrows drawn on the road, and changing colors of traffic lights. Stop signs and road markings have also been accurately displayed in recent updates. In terms of performance, Hamel said on Twitter that features such as Navigate on Autopilot have been improved.

Unlike self-driving companies like Waymo and Cruise, Tesla plans not to use LiDAR. Musk explained that lidar is useful in space, but it seems quite stupid to use it in ordinary cars. In his opinion, just a camera and a neural network are enough to teach a fleet how to drive autonomously.

Tesla is always working to improve its driver-assistance software Autopilot and fully-autonomous driving kits with data collected by the company's fleet. This data is then fed into the company's neural network, which improves upon the input data.

In the past, Tesla has been silent on the official functions of its neural network until a few days ago, Tesla CEO Elon Musk stated on Twitter that an upcoming software update version 2019.40.50 will be released It will include a preview version of the Fully Automated Driving Kit.

This update adds many new features, such as Driver Profiles, Tesla Arcade Games, Trax, and Tesla Theater. But in addition to these improvements, this update also gives people a "first look" at fully autonomous driving.

Of course, Tesla's new visualization capabilities only show the tip of the iceberg for the company's complete autonomous driving solution. The latest visual images released to the company's fleet indicate that Tesla's improvements to its fully autonomous driving capabilities have yielded great results over the past few months.

However, seeing Tesla's electric car recognize objects that were originally ubiquitous on the road suggests that Autopilot and fully autonomous driving are on the right path to true autonomous driving. After all, these visible things are already impressive, and the back-end technology supporting this system may be more advanced. (Little)

Source: Netease Technology Report, translated by Google Translate

Statement: this information is reprinted from authoritative news media. Reprinted for the purpose of transmitting more information and academic exchange, it is not used for commercial purposes, and does not mean to agree with its views or confirm its description. The content of this article is for reference only. If you violate the rights and interests of a third party, please contact us and we will deal with it as soon as possible.