I don't think this has anything to do with XSLT, if you let an HTML parser parse markup like the one you showed then it parses into an HTML object element containing an HTML img element.
document.getElementById('test').innerHTML = `<object><image><\/image><\/object>`;
console.log(document.getElementById('test'));
console.log(document.getElementById('test').outerHTML);<div id="test"></div>So don't expect to be able to parse and preserve XML by parsing it as HTML5.
You can store XML in an HTML script element and read out its .text property:
var scriptElement = document.getElementById('test');
console.log(scriptElement.text);
var xmlDoc = new DOMParser().parseFromString(scriptElement.text, 'application/xml');
console.log(xmlDoc);<script id="test" type="application/xml"><object><image></image></object></script>