跳转到主要内容
由 winniewei 提交于

<p><em>边缘</em><em>A</em><em>I</em><em>助力</em><em>智能</em><em>楼宇实现数字化人流量</em><em>监测</em></p>

<p><span>横跨多重电子应用领域</span><span>、</span><span>全球领先的半导体供应商</span><strong>意法半导体</strong><span>(STMicroelectronics,</span><span>简称</span><span>ST；</span><span>纽约证券交易所代码</span><span>:STM)<span>&nbsp;</span></span><span>与</span><span>能源管理和</span><span>工业</span><span>自动化</span><span>数字化转型</span><span>的</span><span>市场</span><span>领导者</span><strong>施耐德电气</strong><strong><span>&nbsp;</span></strong><span>（</span><span>Schneider Electric</span><span>）</span><span>联合</span><span>推出</span><span>一</span><span>款</span><span>物联网传感器原型</span><span>。通过</span><span>监测</span><span>建筑物的</span><span>居住率</span><span>和使用率，该解决方案可以实现新型物业管理服务，提高</span><span>楼宇的能源管理</span><span>效率。</span></p>

<p><span>两家公司合作</span><span>，</span><span>在</span><span>高性能人数</span><span>统计</span><span>传感器</span><span>内集成</span><span>人工智能（AI）</span><span>技术</span><span>，</span><span>以</span><span>克服在有多个入口的</span><span>大型场所内</span><span>监</span><span>测</span><span>人流量</span><span>的挑战。在</span><span>2020</span><span>年</span><span>11</span><span>月</span><span>19</span><span>日</span><span>IoT</span><span>&amp;</span><span>5G</span><span>研讨会期间，</span><span>施耐德电气将</span><span>以嘉宾身份参加</span><a href="https://www.st.com/content/st_com/en/st-live-days.html?icmp=tt18725_gl_… Live Days</span></a><span>活动，</span><span>演示</span><span>这款物联网</span><span>传感器</span><span>解决方案的功能</span><span>。</span></p>

<p><span>随着</span><span>楼宇及人居</span><span>管理</span><span>数字化</span><span>的提升</span><span>，施耐德致力于成为客户可持续发展和</span><span>提高能源</span><span>效率</span><span>的</span><span>数字合作伙伴，</span><span>为客户</span><span>提供</span><span>有新意、有</span><span>价值的</span><span>洞见</span><span>，如</span><span>，</span><span>协助智能</span><span>楼宇</span><span>管理</span><span>的排队</span><span>监控</span><span>，</span><span>同时</span><span>在设计上</span><span>充分</span><span>尊重</span><span>个人</span><span>隐私。</span><span>该</span><span>先进的</span><span>物联网</span><span>传感器</span><span>解决方案将</span><span>意法半导体</span><span>AI</span><span>产品部</span><span>的</span><span>先进专业知识</span><span>与</span><span>施耐德深厚的传感器应用</span><span>专长相结合</span><span>，将高性能的</span><span>物体</span><span>检测神经网络嵌入到</span><span>经济型</span><span>微控制器(MCU)</span><span>中</span><span>。</span></p>

<p><span><span>施耐德电气通过使用</span></span><span><strong>STM32Cube.AI</strong></span><span><span>工具链提高了设计</span></span><span><span>效率</span></span><span><span>，该工具链具有成熟的功能，可为各种</span></span><a href="https://www.st.com/en/microcontrollers-microprocessors/stm32-32-bit-arm… MCU</span></span></a><span><span>开发AI应用。</span></span><a href="https://www.st.com/content/st_com/en/stm32cube-ecosystem.html"><span><s…;、</span></span><span><span>先进</span></span><span><span>技术</span></span><span><span>和易用性</span></span><span><span>为</span></span><span><span>施耐德电气</span></span><span><span>提供了</span></span><span><span>宝贵的硬件设计灵活性和效率。</span></span></p>

<p><span><span>该</span></span><span><span>人数统计原型传感器</span></span><span><span>整合</span></span><span><span>了</span></span><span><span>施耐德电气</span></span><span><span>独有的</span></span><span><span>超低功耗</span></span><span><span>LYNRED ThermEye</span></span><span><span>TM</span></span><span><span>系列热</span></span><span><span>成像传感器</span></span><span><span>，</span></span><span><span>意法半导体</span></span><span><span>最近推出的高性能</span></span><a href="https://www.st.com/en/microcontrollers-microprocessors/stm32h723-733.ht… MCU</span></span></a><span><span>，</span></span><span><span>以及</span></span><span><span>基于Yolo的神经网络模型。</span></span></p>

<p><span><strong>施耐德电气物联网传感器</strong></span><span><strong>项目</strong></span><span><strong>经理</strong></span><span><strong>Maxime Loidreau</strong></span><span><span>表示：“这项</span></span><span><span>前景广阔</span></span><span><span>的技术</span></span><span><span>解锁</span></span><span><span>一</span></span><span><span>项</span></span><span><span>新的</span></span><span><span>楼宇居住率</span></span><span><span>监测</span></span><span><span>和人</span></span><span><span>流量</span></span><span><span>统计方案，</span></span><span><span>适用于监测</span></span><span><span>排队</span></span><span><span>、</span></span><span><span>建筑使用</span></span><span><span>率、</span></span><span><span>社交距离</span></span><span><span>等</span></span><span><span>多种</span></span><span><span>应用场景</span></span><span><span>。我们与</span></span><span><span>ST</span></span><span><span>合作</span></span><span><span>开发</span></span><span><span>的创新</span></span><span><span>的</span></span><span><span>演示</span></span><span><span>原型不仅</span></span><span><span>适用于</span></span><span><span>酒店</span></span><span><span>、</span></span><span><span>办公室</span></span><span><span>、</span></span><span><span>百货</span></span><span><span>零售</span></span><span><span>等场所，也适用任何需要掌握人流量及</span></span><span><span>空间</span></span><span><span>使用率等情况</span></span><span><span>的</span></span><span><span>任何场景</span></span><span><span>。这将重新定义未来建筑</span></span><span><span>！”</span></span></p>

<p><span><strong>意法半导体的</strong></span><span><strong>AI</strong></span><span><strong>解决方案业务线经理</strong></span><span><strong>Miguel Castro</strong></span><span><span>补充说：“该项目</span></span><span><span>证明</span></span><span><span>深度学习</span></span><span><span>能够提升</span></span><span><span>嵌入式数据处理性能，展示了如何在</span></span><span><span>基于经济型</span></span><span><span>微控制器的</span></span><span><span>系统</span></span><span><span>平台上</span></span><span><span>运行</span></span><span><span>高价值</span></span><span><span>的</span></span><span><span>应用</span></span><span><span>软件</span></span><span><span>。我们的STM32Cube.AI生态系统使用户能够在</span></span><span><span>很短的</span></span><span><span>时间内</span></span><span><span>开发出</span></span><span><span>灵活的解决方案</span></span><span><span>，</span></span><span><span>利用我们技术团队的支持</span></span><span><span>服务</span></span><span><span>克服工程挑战，客户可以享受更高的</span></span><span><span>设计效率</span></span><span><span>。</span></span><span><span>”</span></span></p>

<p><span><strong>技术详情</strong></span></p>

<p><span><span>STM32 AI</span></span><span><span>生态系统提供了在</span></span><span><span>STM32 MCU</span></span><span><span>上运行神经网络所需的基本模块，可以实现经济、节能的解决方案，原生支持各种深度学习框架</span></span><span><span>，</span></span><span><span>例如：</span></span><span><span>Keras</span></span><span><span>、</span></span><span><span>TensorFlow™Lite</span></span><span><span>和</span></span><span><span>ONNX</span></span><span><span>交换格式。</span></span></p>

<p><span><span>在这个开发生态系统中，</span></span><span><span>X-CUBE-AI</span></span><span><span>软件扩展包扩大了</span></span><span><span>STM32CubeMX</span></span><span><span>初始化工具的功能，</span></span><span><span>可</span></span><span><span>以自动转换预先训练的神经网络，为目标</span></span><span><span>MCU</span></span><span><span>生成优化的软件库，并集成到用户的项目中。此外，还有很多其他的自动化功能，把开发者从繁重的开发任务中解放出来，例如，可以验证神经网络模型并测量</span></span><span><span>STM32 MCU</span></span><span><span>的性能，而无需人工开发</span></span><span><span>C</span></span><span><span>代码。</span></span></p>

<p><span><span>意法半导体的软件开发生态系统支持的普通</span></span><span><span>DNN</span></span><span><span>方案已经映射到</span></span><span><span>STM32</span></span><span><span>丰富的产品组合内，使用户可以有效地复制开发成果，为多个市场创建产品。在</span></span><span><span>ST Live Days</span></span><span><span>上运行演示应用的</span></span><span><span>STM32H723 MCU</span></span><span><span>完全胜任运行</span></span><span><span>AI</span></span><span><span>应用</span></span><span><span>，因为这款产品的内核</span></span><span><span>性能优异</span></span><span><span>，闪存容量</span></span><span><span>高达</span></span><span><span>1Mbyte</span></span><span><span>，提供</span></span><span><span>高速片外存储器接口</span></span><span><span>，</span></span><span><span>以及各种传感器接口。</span></span></p>

<p><span><span>了解</span></span><span><span>STM32Cube.AI</span></span><span><span>更多信息</span></span><span><span>，请</span></span><span><span>访问</span></span><a href="http://www.st.com/STM32CubeAI"><span><span>www.st.com/STM32CubeAI</span…;

<p><span><span>了解</span></span><span><span>STM32H7 MCU</span></span><span><span>详情，请访问：</span></span><a href="https://www.st.com/en/microcontrollers-microprocessors/stm32h7-series.h…;

<p><span><span>了解如何在</span></span><span><span>STM32</span></span><span><span>微控制器和应用处理器上运行边缘</span></span><span><span>AI</span></span><span><span>应用</span></span><span><span>，请</span></span><span><span>联系我们</span></span><span><span>：</span></span><a href="mailto:edge.ai@st.com"><span><span>edge.ai@st.com</span></span></a></p&…;

<p><strong><span>关于意法半导体</span></strong></p>

<p><span>意法半导体拥有46,000名半导体技术、产品和方案的创新者和创造者，掌握半导体供应链和最先进的制造设备。作为一家独立的半导体设备制造商，意法半导体与十万余客户、上千合作伙伴一起研发产品和解决方案，共同构建生态系统，帮助他们更好地应对各种挑战和新机遇，满足世界对可持续发展的更高需求。意法半导体的技术让人们的出行更智能，电力和能源管理更高效，物联网和5G技术应用更广泛。详情请浏览意法半导体公司网站：</span><a href="http://www.st.com/"><span>www.st.com</span></a><span&gt;。</span></p>