为什么要在自动化项目中实施 3D 视觉？

今天的自动化项目越来越先进。 3D 视觉是相关改进的有力示例。这就是为什么它对于帮助项目达到新的高度如此不可或缺，并了解该技术多年来如何发展以解决以前的缺点并增强当前和未来应用程序的多功能性。

3D 视觉如何改进自动化

该技术允许应用程序检测物体，而不管它们的位置如何。无论物品是堆叠的、装在箱子里的还是以其他方式排列的，这些解决方案都通过使其越来越适用于广泛的场景来推动工业自动化向前发展。这些系统中的许多系统都被设置为使它们的行为类似于人眼，例如通过包含两个镜头来处理传入的信息。

真实世界的环境通常不符合整洁的参数。它可能具有不同的纹理、不均匀的对象高度以及其他很少或永远不会在不同时刻保持静态的元素。幸运的是，3D 传感适应了这些差异。它使机器人能够适应非结构​​化环境并更加擅长新任务。

计算机视觉系统区分机器人、人和其他物体的能力也很重要，因为员工通常与各种类型的机器人技术一起工作，以更快地完成工作并减少错误。当自动化机器能够正确感知拥挤环境的具体情况时，这种能力会使其整体更加安全。

随着自动化应用程序变得越来越多样化和先进，人们将继续了解为什么三维计算机视觉对于使它们按预期工作至关重要。公司领导通常部署智能设备以降低错误率并提高生产率。检测 3D 对象的机器视觉可以提供这些优势以及更多优势。

部署工业 3D 视觉的常见机器人应用

许多最先进的机器人都具有三维视觉功能，可以增强它们在执行某些任务时的感知能力。以下是一些潜在应用的示例。

制造

许多工厂使用自动化移动机器人 (AMR) 来应对劳动力短缺的挑战，并减少团队成员从事繁重重复性工作的需要。 AMR 通常具有机载视觉功能，可以让机器通过识别障碍物并绕过障碍物找到清晰的路径。

医疗保健

改进的机器感知对于需要非凡精度的应用（例如机器人手术）也至关重要。一家公司已将该技术应用于大脑和脊柱手术，使外科医生能够看到必要的区域，而无需让患者暴露在辐射下。

该系统还可以通过创建与手术相关的身体部位的三维图像画布来改进手术计划。

农业

如果应用于大规模农业，这项技术对工业自动化的好处也会变得显而易见。一些机器人可以只专注于成熟的农产品，使整个收获过程更加有效。

研究人员还建议，3D 机器视觉解决方案可以帮助奶农评估牛的身体成分。进行此类评估对于保持奶牛泌乳、帮助它们怀孕和减少代谢疾病至关重要。

食品和饮料

食品和饮料公司使用计算机视觉检测系统来检查从酸奶到比萨饼面团的所有东西。此类应用可以大大加快质量控制。

然而，计划使用它的人必须意识到来自太阳或头顶人工光源的环境光可能会影响这些检查的质量和一致性。因此，了解如何维护环境以获得最佳结果至关重要。

物流

考虑到包裹的各种形状和大小，物流公司开始利用 3D 机器视觉的力量来提高生产力也就不足为奇了。 DHL 就是这样一个品牌。

它最近投资了用于卸垛的机器人。除了改进工作流程输出外，决策者还相信机器人应用将带来更好的工人安全。

零售

Not so long ago, retail workers performed manual stock checks to gauge how fast certain products were selling and when to replenish the shelves. That still happens to some extent, but many retail brands now let robots equipped with 3D machine vision do that task. At some grocery retailers, including ShopRite, a bot roams the aisles up to three times per day to check product statuses. The high-tech machines can also detect when items are in the wrong places.

Advancements in 3D vision that tackle automation challenges

The above examples show how 3D sensing can help various industries get impressive results and target pain points. Improvements in the technology have also made it easier to deploy a robot for diverse tasks that were too complicated for 2D systems to do well. Examples include picking up parts from piles.

Three-dimensional vision systems are often the more appropriate choice because 2D vision systems do not provide depth information and as such, they are only suitable for simple applications that require the capture of objects in the X and Y-axis. Examples of such applications include barcode reading, character recognition, dimension checking, or label verification. 3D machine vision systems enable much more complex robotic tasks because they provide 3D point clouds with precise X, Y, and Z coordinates.

Challenges can also arise when using robots for bin-picking tasks due to the potential variations in product placement within those containers. Sometimes, the item positions are structured, such as if they’re stacked or lined up in rows. However, there are other cases where products overlap inside bins or are even entangled. Luckily, the most advanced machine vision systems can handle such scenarios without issues.

Better industrial automation made possible with Parallel Structured Light technology

Photoneo also has a novel Parallel Structured Light technology that powers its 3D camera, known as MotionCam-3D. Traditional area-sensing technologies could not handle the capture and scanning of moving 3D scenes without unwanted blur and artifacts.

It relies on a snapshot system that constructs multiple virtual images within a single exposure window. This approach gives high-resolution, accurate results. MotionCam-3D can scan dynamic scenes moving up to 144 km/hour. Alternatively, companies can use it for static scenes. Now, the capabilities of the camera have been extended to provide also color information.

Three-dimensional computer vision supports automation

These examples show why decision-makers in a wide range of industries should strongly consider exploring how 3D vision could support their intentions to automate particular processes. Implementing this technology takes time and effort, but it’s usually worthwhile.