New optical sensor powers robotics and image recognition AI

Oregon State University in the United States has made significant progress in the research and development of new optical sensors. They have developed a new optical sensor that more realistically mimics the human eye's ability to perceive changes in objects. The achievement is expected to lead to major breakthroughs in areas such as image recognition, robotics and artificial intelligence. An optical sensor is a sensor that measures based on optical principles. It has many advantages, such as non-contact and non-destructive measurement, almost no interference, high-speed transmission, and telemetry, remote control, etc. It mainly includes optical sensors and instruments such as general optical measuring instruments, laser interference type, grating, encoder and optical fiber type. It is mainly designed to detect the presence of objects, or to perform motion detection in various industrial, automotive, electronic products and retail automation. Current information processing algorithms and architectures are becoming more and more like the human brain, but the way information is received is still designed for traditional computers. To reach its full potential, be more like the human brain“think”of computers need to be more like the human eye“Look”image sensor. Optical sensors mainly include: optical image sensors, transmissive optical sensors, optical measurement sensors, optical mouse sensors, reflective optical sensors, etc. Traditional sensing technologies, such as the chips in digital cameras and smartphones, are better suited for sequential processing. Each sensor produces a signal whose amplitude varies with the intensity of light it receives, which means that a static image results in a relatively constant output voltage from the sensor. The unique optoelectronic properties of perovskites are exploited in a novel retinal morphological sensor. Perovskites act as capacitors in ultrathin layers just a few hundred nanometers thick, which change from electrical insulators to conductors when exposed to light. The sensor thus remains relatively quiet under static conditions, recording a short, sharp signal when a change in illumination is detected, then quickly returning to its baseline state. The new sensor is also a perfect match for neuromorphic computers. Unlike traditional computers, neuromorphic computers are massively parallel networks that mimic the human brain, powering the next generation of artificial intelligence used in self-driving cars, robotics and advanced image recognition.